Alex Merced's Lakehouse Blog

Skip to Part 5: Going Advanced: Open Source, Local Models, and Agent Tools

Getting Started with AI for Free: Every Tool Google Gives You at No Cost

Alex Merced — Mon, 01 Jun 2026 09:00:00 GMT

Most people think you need to pay $20 or $200 a month to get value from AI. That is not true. If you have a Google account which is free and most people do you already have access to a surprisingly powerful set of AI tools. No credit card required. No upgrade needed.

This is Part 2 of "Catching Up with Using AI for All Levels." In Part 1, we covered what AI actually is under the hood: prediction engines built on vectors and transformers, not thinking machines. Now we get practical. This post walks through every AI tool Google provides at no cost, what each one does well, and concrete ways to use them in your daily life. Part 3 will cover ChatGPT and Claude for when you are ready to pay. Part 4 covers specialized creative tools. Part 5 goes deep into open source and local models.

If you missed Part 1: What AI Is and Isnt, start there for the foundation.

Continue to Part 3: ChatGPT and Claude Deep Dive

Skip to Part 5: Going Advanced: Open Source, Local Models, and Agent Tools

The Google AI Ecosystem: A Quick Overview

Google has been an AI company longer than most people realize. They acquired DeepMind in 2014. They pioneered the transformer architecture that powers every major LLM today. They have the largest AI research organization in the world. And they have been quietly integrating AI into their free products for years.

The 2026 free tier is more generous than any other major AI company offers. Here is what you get at no cost:

Gemini (the chatbot). Access to Google's Gemini model through gemini.google.com, the Gemini mobile app, and integrations across Google services. The free tier includes text chat, voice input, image analysis, and limited file uploads.

Gemini in Gmail and Google Docs. The "Help me write" feature that drafts emails, summarizes threads, and polishes text. This is built into your existing free Google account.

NotebookLM. Google's AI research assistant. Upload PDFs, YouTube videos, websites, and audio files. Get summaries, ask questions about your sources, and generate audio overviews. The free tier gives you 100 notebooks with 50 sources per notebook.

Google AI Studio. A web based playground for experimenting with Gemini models directly. You get API access with free rate limits for prototyping and testing.

Gemini CLI. Run Gemini models from your terminal or scripts. Up to 1,000 requests per day with a Google account.

Google Antigravity. Google's agent first IDE is free during public preview. Unlimited tab completions and command requests with weekly rate limits for fairness.

Gemini Code Assist for Individuals. Free code completion and generation in VS Code, JetBrains, and other IDEs. Up to 180,000 code completions per month with no credit card.

Classic AI APIs. Translation (500,000 characters/month), Speech to Text (60 minutes/month), Text to Speech (4 million standard plus 1 million WaveNet characters/month), Cloud Vision (1,000 units/month), Natural Language API (5,000 units/month), and Video Intelligence (1,000 minutes/month). All free with no expiration.

Let us go through each one in detail with real productivity examples.

Gemini: Your Free AI Assistant

Gemini is Google's direct competitor to ChatGPT and Claude. The free tier is more limited than the paid versions, but it still covers a lot of ground.

What You Get for Free

The free tier gives you access to Gemini Pro models with text chat, voice input, and image understanding. You can upload images and ask questions about them. You can have it read and summarize PDFs and documents up to certain size limits. You get the Gemini mobile app on iOS and Android with voice conversations.

What you do not get: access to Gemini Ultra or the latest flagship models, Gemini Spark for real time web searching, larger file uploads, priority access during high traffic, or integration with Google Calendar and Tasks. Those require a Google AI Premium ($20/month) or AI Ultra ($100/month) subscription.

Daily Productivity Uses for Free Gemini

Email drafting. Open Gmail, click the "Help me write" button, and describe the email you want to send. "Draft a short email to my team about rescheduling tomorrow's standup to 10:30 AM." It writes the email in your voice. You review and send. This alone saves several minutes per email, and if you send ten emails a day, that adds up to nearly an hour a week.

One trick that regular users discover is that you can train Gemini to match your writing style. Before asking it to draft an email, paste in three emails you wrote recently and say "Match this style." The model picks up on your sentence length, formality level, and common phrases. Over time, the drafts need less editing.

Document polishing. In Google Docs, use "Help me write" to rewrite a paragraph, make it more concise, or change the tone. This works on any document in your Google Drive. The most practical use is the "Make shorter" and "Make more formal" buttons. Your first draft goes into the document. You mark the rough paragraph and ask for a polished version. Review the suggestions, accept the ones that work, and move on.

Quick research. Ask Gemini to explain a concept, summarize a news article, or give you an overview of a topic. It has access to current information through Google Search integration on the free tier. The research quality depends on how specific you are. "Explain how mortgages work" gives you a generic overview. "Compare 30 year fixed rate mortgages to 5 year adjustable rate mortgages for someone planning to stay in their home for 3 years" gives you a focused answer that helps with a real decision.

Image analysis. Take a photo of a whiteboard after a meeting and ask Gemini to transcribe and organize the notes. Take a photo of a nutrition label and ask for a summary of the ingredients. Take a photo of a plant with brown leaves and ask what might be wrong. The vision capabilities work across photos, screenshots, and scanned documents.

Brainstorming. "Give me ten ideas for team building activities that cost less than $50." "Suggest three different ways to structure this presentation." "Help me come up with a name for my new side project." The key to good brainstorming with AI is to ask for quantity first, then refine. Ask for 20 ideas, not 5. Many will be unusable, but the few good ones make the exercise worthwhile.

Translation. "Translate this email from Spanish to English." "What does this French menu item mean?" Gemini handles over 100 languages. The translation quality is competitive with dedicated translation tools for common language pairs like Spanish English, French English, and Chinese English.

Content repurposing. Paste a long memo into Gemini and ask for three versions: a one paragraph summary, a bullet point list of key takeaways, and a social media post announcing the main finding. This turns one piece of content into three formats in under a minute.

Where Free Gemini Falls Short

The free tier has lower rate limits than paid. During peak usage, you might get slower responses. You cannot use it for automated workflows or API calls at scale. The context window is smaller than the paid version, meaning it handles shorter documents. You also cannot connect it to your Google Calendar or Tasks to manage your schedule.

If you find yourself hitting these limits regularly, the $20/month Google AI Premium plan removes most of them and adds integration with Google apps. But for casual daily use, the free tier is genuinely useful.

NotebookLM: Your Personal Research Assistant

NotebookLM is arguably Google's most underrated free AI tool. It is a research assistant that works with your own documents. You upload sources, and the AI answers questions based only on those sources. It does not go off script. It does not make up facts from its general training data. It works from your materials.

What Makes NotebookLM Different

Unlike Gemini or ChatGPT, NotebookLM has a source grounded architecture. When you ask a question, it searches only the sources you uploaded and generates answers from that content. Every answer includes citations showing exactly which source and passage it used. This makes it much more reliable for fact based work.

The free tier allows 100 notebooks with up to 50 sources each. Each source can contain up to 500,000 words. That is roughly 1,000 pages per source, or 50,000 pages per notebook. You are not going to hit these limits with normal use.

NotebookLM supports PDFs, Google Docs, websites (paste a URL), YouTube videos (paste a link), and audio files. The audio feature is particularly notable. You can upload a recording of a meeting or lecture, and NotebookLM will transcribe it and let you ask questions about the content.

Audio Overviews: The Hidden Gem

NotebookLM can generate an Audio Overview a podcast style conversation between two AI hosts that discusses your sources. This is surprisingly good. The hosts summarize the material, make connections between sources, and ask each other questions. It sounds like a real podcast discussion about your research.

This feature is useful for processing long documents while doing other things. Upload a 50 page report, generate an Audio Overview, and listen to the summary during your commute. You still need to read the original for details, but the overview gives you the big picture.

Daily Productivity Uses for NotebookLM

Meeting prep. Upload the agenda, previous meeting notes, and relevant documents. Ask NotebookLM to summarize the key decisions from the last meeting and list open action items. "Based on these notes, what topics are likely to come up in today's meeting?" The source grounding means every claim in the answer can be traced back to a specific document.

Research synthesis. Upload five articles on the same topic. Ask NotebookLM to identify areas of agreement and disagreement between the sources. Ask it to extract all statistics mentioned across the sources into a table. Ask it to create a timeline of events described across the documents. This turns an hour of reading into five minutes of review.

Learning a new topic. Upload a textbook chapter, a video transcript from a course, and a few relevant articles. Use NotebookLM as a tutor. "Explain this concept as if I were a beginner." "Give me practice questions about chapter 3." "Create a study guide with the ten most important concepts from these sources." The fact that answers are source grounded means you can dig into the original material when something is unclear.

Job applications. Upload the job description and your resume. Ask NotebookLM to identify gaps between your experience and the requirements, then suggest how to frame your experience for the role. "What keywords from the job description should I include in my cover letter?" Upload the company's About page and recent press releases, then ask for insights to use in an interview.

Legal and compliance reading. Upload contracts or policy documents. Ask NotebookLM to extract key obligations, deadlines, and restrictions. "List all dates mentioned in this contract and what happens on each date." "What are the termination conditions?" The source grounding means you can trust the answers more than a general chatbot, but you should still read the original document for anything important.

Client research. Before a client meeting, upload their website, recent press releases, annual report, and any previous correspondence. Ask NotebookLM to create a briefing document: company overview, recent developments, likely priorities, and potential talking points. The briefing takes minutes to generate instead of hours of manual research.

Content creation workflow. Upload your notes and research into NotebookLM. Ask it to organize the information into an outline. Export the outline to Google Docs and use Gemini to expand each section. Use the Audio Overview feature to generate a podcast style summary of the finished piece. One flow from research to outline to draft to audio, all within Google's free tools.

Google AI Studio: Where to Experiment

Google AI Studio is a web based environment for testing Gemini models directly. It is aimed at developers and curious users who want to understand how models behave.

The free tier gives you API access with rate limits that are generous enough for learning and prototyping. You can try different models, adjust parameters like temperature and top_p, and see how the model responds to different prompts. The interface shows you the token usage, latency, and safety attributes for each request.

What You Can Do with AI Studio

You do not need to be a developer to use AI Studio. The interface is point and click. You type a prompt, pick a model, and see the output. But the real power comes from being able to tune parameters and see the effects.

Try setting temperature to 0 and asking the same question three times. You get the same answer each time. Then set temperature to 1 and ask the same question. The answers vary. This is a practical way to understand the prediction mechanism we covered in Part 1.

AI Studio also includes system instructions, a feature that lets you set the behavior and persona of the model before the conversation starts. "You are a Spanish tutor. Keep responses under 100 words. Include one grammar tip each response." This changes how the model behaves without you having to repeat instructions in every message.

Why You Should Try AI Studio

Even if you never write code, spending 30 minutes in AI Studio will improve how you use every other AI tool. You will see how small changes in prompt wording change the output. You will understand why the same model gives different answers to the same question. You will learn about temperature, top_p, and system instructions concepts that apply across ChatGPT, Claude, and every other AI service.

Gemini Code Assist: AI for Your Coding (Even Beginners)

Gemini Code Assist for Individuals is free, requires no credit card, and works with VS Code, JetBrains IDEs, and other popular editors. It provides real time code completion, code generation from comments, and debugging assistance.

Wait, I Do Not Code

Fair point. If you are not a developer, this section might not apply directly. But keep reading for two reasons.

First, you might encounter situations where a small script would save you hours of manual work. With Code Assist, you can describe what you want in plain English: "Write a Python script that renames all files in a folder to match a pattern." It generates the code. You run it. Problem solved without learning to code.

Second, you can use Gemini CLI even without being a programmer. It lets you run Gemini from the command line. The command gemini ask "What is the weather in Chicago today?" works on any system with the CLI installed. No coding required.

The Classic Google AI APIs: Free and Overlooked

Google Cloud offers free tiers for a dozen AI services that reset monthly and never expire. Most people do not know these exist. Here is what stands out.

Translation API

500,000 characters free per month. That is roughly 80,000 words of translation every month at no cost. You can use it through Google Translate on the web, which is free anyway, but the API lets developers integrate translation into applications. For personal use, the Google Translate app and website already give you the same capability.

Speech to Text

60 free minutes of transcription per month. Upload an audio file and get a timestamped transcript. This is useful for transcribing meeting recordings, lectures, interviews, or voice memos. The accuracy is good for clear audio and degrades with background noise or heavy accents.

Text to Speech

4 million characters of standard voices plus 1 million characters of WaveNet (high quality) voices per month. Convert any text into natural sounding speech. Useful for creating audio versions of documents, accessibility, or language learning.

Cloud Vision

1,000 free units per month. Upload an image and get back detected objects, faces, text, and landmarks. This is the technology behind Google Lens. On the free tier, you can use it through the Google Cloud Console web interface without writing any code.

Natural Language API

5,000 units per month for entity analysis, sentiment analysis, and syntax analysis. Paste in a block of text and get back the detected people, places, organizations, and the overall sentiment. Useful for analyzing customer feedback or social media mentions.

Video Intelligence

1,000 free minutes per month. Upload a video and get shot detection, label detection, and explicit content detection. This is more specialized, but if you work with video content, the free tier gives you significant processing capacity.

Google Antigravity: The New Player

Google Antigravity is Google's agent first IDE announced at Google I/O 2026. It is free during public preview. Think of it as AI development environment that goes beyond code completion. It integrates Gemini directly into the development workflow with tab completions, command requests, and agentic assistance.

For non developers, Antigravity is less relevant today. But if the preview hints at the direction Google is heading, we can expect more agent driven tools that handle complex multi step tasks on your behalf. Keep an eye on this one.

Practical Workflows: Putting Free AI to Work

Here are complete workflows that use only free Google AI tools.

Morning Information Intake

Wake up. Open Gemini on your phone. Say "What happened overnight that I should know about?" Gemini summarizes news based on your interests. If you have a Google Nest Hub, Gemini is built into the smart display and can give you a verbal briefing while you make coffee.

Meeting Follow Up

After a meeting, upload the recording to NotebookLM. It transcribes the audio automatically. Ask NotebookLM: "Summarize the key decisions from this meeting. List the action items with owners. Identify any dates mentioned." Export the summary as a document and share it with attendees.

Research for a Purchase Decision

You are buying a new laptop. Collect links to five review sites, two comparison articles, and the official specs sheet. Paste the URLs into NotebookLM sources. Ask: "Compare the MacBook Air M4 and the Dell XPS 16 on build quality, performance, battery life, and price. Create a table." The answers are sourced from your documents, not from the model's general knowledge.

Writing Assistance

Open a Google Doc. Write a rough draft of your email or document. Use the "Help me write" sidebar to polish the language, shorten it, or adjust the tone. Then use Gemini to check: "Read this and tell me if there are any unclear sentences."

Language Learning

Use Gemini for conversation practice in your target language. "Let us practice Spanish. Ask me questions about my weekend and correct my grammar mistakes." Use the Text to Speech API through the Google Cloud Console to hear native pronunciation. Upload articles in your target language to NotebookLM and ask for translations of specific passages.

Workflow Automation (Low Code)

Install the Gemini CLI. Create a simple script that checks your Gmail for specific types of messages and summarizes them. The free tier handles 1,000 requests per day, which is more than enough for personal use. You do not need to be a programmer to copy a script from the documentation and run it.

Here is a concrete example. Set up a recurring task that uses Gemini CLI to read the headlines from a few news RSS feeds and summarize them into a morning briefing. Save the output to a Google Doc. The whole pipeline runs automatically without you touching anything.

Teaching and Tutoring

Use Gemini as a practice partner for learning. Studying for a certification exam? Ask Gemini to quiz you on each topic. Learning a new language? Have Gemini conduct a conversation and correct your mistakes. Preparing for a presentation? Practice your talking points and ask Gemini to critique your logic and suggest better examples.

The key is to treat AI as a patient, always available practice partner that never gets bored or impatient. It will quiz you on the same topic fifty times if that is what you need.

Travel Planning

Use Gemini to research destinations, compare flight options, and create itinerary drafts. "Plan a 5 day trip to Lisbon focused on food and history. Include free activities for two of the days." Then use NotebookLM to store all your research: hotel confirmations, restaurant recommendations, museum hours, and transportation options. One notebook per trip keeps everything organized and searchable.

Google Photos and Google Lens: AI You Are Already Using

Two more free AI tools worth mentioning because you probably use them without thinking about it.

Google Photos uses AI for automatic categorization. You can search "dogs" and it finds every photo containing a dog. You can search "beach" and it finds vacation photos. You can search "birthday cake" and it surfaces party photos. The AI identifies faces, objects, locations, and even specific events in your photo library. All of this runs for free on your Google account with 15GB of storage.

Google Lens is image recognition built into your phone's camera. Point it at a landmark and it identifies the building. Point it at a plant and it tells you the species. Point it at a document and it extracts the text. Point it at a product and it finds prices online. Lens uses the same underlying vision AI as Cloud Vision but through a free consumer interface. It is one of the most practical AI tools you already have in your pocket.

When Free Is Not Enough

The free tier is generous, but it has limits. Here are the signs that you might benefit from a paid subscription.

You hit rate limits regularly. Gemini gets slow during peak hours. You need to process very long documents frequently. You want to connect AI to your calendar, tasks, and other personal data. You need priority access for time sensitive work.

Google's paid tiers start at $20/month for Google AI Premium (formerly Google One AI Premium), which gives you Gemini Ultra access, larger context windows, Google Workspace integration, and priority access. The $100/month AI Ultra plan adds higher usage limits and access to the latest models.

But before you pay, ask yourself whether the free tools actually solve your problems. Many people pay for AI subscriptions out of FOMO and end up using the same features they had for free. Start with the free tier. Use it for a month. If you hit real limitations, then consider upgrading.

What Comes Next

This post covered more than a dozen free AI tools available through your Google account. The most useful for daily productivity are Gemini (chat and writing), NotebookLM (research), AI Studio (learning and experimentation), and the classic APIs for translation, transcription, and vision.

Part 3 of this series covers ChatGPT and Claude, the two main competitors to Google's offerings. You will learn what you get at each paid tier, how to use desktop apps and advanced features like Clips and Dispatch, and whether the paid plans are worth it for your use case.

Continue to Part 3: ChatGPT and Claude Deep Dive

Skip to Part 5: Going Advanced: Open Source, Local Models, and Agent Tools

ChatGPT and Claude: Which AI Service Should You Pay For

Alex Merced — Mon, 01 Jun 2026 09:00:00 GMT

Part 2 of this series covered the extensive free AI tools Google offers through your Gmail account. Now we step up to the paid tier. ChatGPT from OpenAI and Claude from Anthropic are the two most popular paid AI assistants in 2026. Between them, they handle the vast majority of AI interactions worldwide.

This is Part 3 of "Catching Up with Using AI for All Levels." If you are just joining, start with Part 1 for the fundamentals of how AI works and Part 2 for the free tools. This post covers the paid side: what each service charges, what you get at each tier, the desktop apps and advanced features, and practical examples of how to use everything for daily productivity.

Skip to Part 5: Going Advanced: Open Source, Local Models, and Agent Tools

The Two Giants: ChatGPT and Claude

OpenAI launched ChatGPT in November 2022 and sparked the current AI boom. Anthropic launched Claude shortly after, positioning it as a safety focused alternative. In 2026, both companies have mature products with overlapping features and distinct strengths.

The choice between them is not about which one is "better." It is about which one fits your specific needs. They are more similar than different at the basic level, but their advanced features diverge significantly.

ChatGPT: Features and Pricing

Free Tier

ChatGPT's free tier is more restrictive than Google's free offerings. You get access to GPT 5.0 Mini with text chat, limited file uploads, and basic image generation through DALL E integration. The free tier uses an older, smaller model and has rate limits that slow down during peak usage.

You can use it for simple tasks: answering questions, drafting short text, basic brainstorming. But the free tier is designed to give you a taste, not a full experience. Most users hit the limits within a few sessions and consider upgrading.

ChatGPT Plus: $20 per Month

The $20 tier is where ChatGPT becomes genuinely useful. Here is what you get.

Access to GPT 5 series models. GPT 5.1 and GPT 5.2 are the standard models available to Plus users. They are significantly more capable than the free tier model: better reasoning, longer context windows, better instruction following. GPT 5.4, the latest flagship model released in early 2026, is also available to Plus users with usage caps. You get a certain number of messages per day on the flagship model before it reverts to 5.2 for the rest of the day.

The practical difference between the models is noticeable. GPT 5.4 handles complex multi step instructions more reliably, writes more coherent long form content, and makes fewer logical errors. For quick questions and simple tasks, 5.2 performs almost as well. The tiered model access means you save your limited flagship messages for the hardest tasks.

Voice conversations. The advanced voice mode on the mobile app supports real time conversation with emotional range. The model detects your tone and adjusts its response accordingly. You can interrupt it mid sentence. It laughs at appropriate moments. The voice quality is good enough for extended conversations, and the latency is low enough that the conversation feels natural.

The voice mode has become a primary interface for many users. They dictate emails, brainstorm ideas verbally, practice presentations, and have the model read documents aloud. It is surprisingly effective for tasks where typing is inconvenient.

Image generation. Built on DALL E 3 and the newer DALL E 4 model. You can generate images from text descriptions, edit existing images, and create variations. The quality is competitive with dedicated image tools for many use cases, though specialized tools like Midjourney still lead for artistic work.

The image generation is deeply integrated into the chat interface. You can generate an image, discuss it with the model, request changes, and iterate without switching applications. This tight feedback loop is more efficient than using separate tools for conversation and image generation.

File uploads and data analysis. Upload PDFs, Word documents, Excel spreadsheets, and images. ChatGPT reads the content and can analyze it. You can ask it to find patterns in a spreadsheet, summarize a long PDF, or extract data from images. This feature works well enough for everyday analysis but hits limits with very large files or complex multi sheet workbooks.

One underappreciated use: upload a messy CSV file and ask ChatGPT to clean it. "Remove duplicate rows. Standardize the date column to ISO format. Flag any rows with missing values." ChatGPT processes the data and gives you a cleaned version. For non technical users, this replaces a dozen spreadsheet formulas.

Web browsing. ChatGPT can search the internet for current information when you enable the browsing feature. It cites its sources. This is useful for research that requires up to date information, though the browsing is slower than using a search engine directly because the model needs to search, read, and synthesize in sequence.

Custom GPTs. Create your own specialized versions of ChatGPT with custom instructions, specific knowledge files, and configured capabilities. The GPT Store offers thousands of community created GPTs for specific tasks: writing assistant, code reviewer, travel planner, fitness coach. Most of these are shallow wrappers, but a well made custom GPT can save significant setup time for recurring tasks.

The best use of Custom GPTs is creating one for your own recurring workflows. A "Meeting Notes GPT" that always formats output the same way. A "Content Repurposer" that turns any document into social posts, a newsletter summary, and a blog outline. The time saved from not repeating instructions adds up.

Memory. ChatGPT Plus includes persistent memory across sessions. You can tell the model facts about yourself once, and it remembers them in future conversations. "I prefer concise answers." "I work in healthcare compliance." "My team has seven people." These details persist and shape future responses without you repeating them.

ChatGPT Pro: $200 per Month

The $200 tier removes most usage caps. You get unlimited access to GPT 5.4, priority during peak times, higher file upload limits, and priority access to new features. This tier is for power users who depend on ChatGPT for their daily work and hit the Plus tier's limits regularly.

The main question to ask yourself: are you hitting the Plus limits more than once a week? If yes, Pro might be worth it. If no, save your money.

Claude: Features and Pricing

Claude Free

Claude's free tier gives you access to Claude Sonnet, the mid tier model. It is more generous than ChatGPT's free tier in terms of model quality but still has rate limits. You can use it for text chat, file uploads (PDFs, images, documents), and basic analysis. The free tier works for light use: quick questions, document summaries, short writing tasks.

Claude Pro: $20 per Month

The Pro tier upgrades you to Claude Opus, Anthropic's best non Max model. Here is what you get.

Claude Opus access. Opus is Anthropic's flagship model, comparable to GPT 5.4 in capability. It excels at reasoning, writing, and analysis. Many users report that Claude Opus produces more naturally flowing, better structured long form writing than ChatGPT. The difference is subjective and task dependent, but it is a real distinction.

200K token context window. Claude's context window has been a differentiator since its early days. 200,000 tokens is roughly 150,000 words, or about 300 pages of text. You can paste an entire novel into a single prompt and ask questions about it. This is useful for analyzing large documents, comparing multiple files, or working with long codebases.

Projects. A Claude feature that lets you organize conversations, documents, and custom instructions into dedicated workspaces. Each project can have its own knowledge base (uploaded documents), custom instructions, and conversation history. This is useful for ongoing work on a specific topic: a project for your book research, another for your job search, another for your side project. The projects feature is one of Claude's strongest organizational tools and something ChatGPT does not directly replicate.

Claude Mobile App. Full featured mobile app with voice input, image upload, and conversation sync across devices. The mobile experience is similar to ChatGPT's.

Claude Code. Claude's coding assistant that runs in your terminal. It can read and write files, execute commands, and manage entire codebases. Claude Code is more integrated into the developer workflow than ChatGPT's code features because it operates directly in your terminal environment. It can create files, run your project's test suite, and fix errors automatically.

Claude Max: $200 per Month

Claude Max is Anthropic's top tier, and this is where Claude pulls ahead of ChatGPT for power users. Here is what makes it different.

5x to 20x higher usage limits. You can use Claude Opus essentially without worrying about rate limits. For heavy daily users, this is the main reason to upgrade. Instead of rationing your messages, you use Claude freely throughout the day for every task, question, and document analysis.

Claude Cowork. This is the desktop agent feature. Claude Cowork runs on your computer and can see your screen, interact with applications, read and write files, and perform multi step tasks. It is like having an assistant that can actually use your computer.

The practical difference from a regular chatbot is enormous. Normal chatbots can only process text you give them. Cowork can see your screen. You can ask it questions about anything visible on your monitor. "Which Slack message from the engineering channel is unresolved?" "What is the stock price shown in the browser tab on the left?" It reads the information directly from your screen.

Cowork can also take actions. "Open the Chrome browser, navigate to my analytics dashboard, take a screenshot of the weekly active users chart, and save it to the Desktop." Claude opens Chrome, types the URL, waits for the page to load, captures the screenshot, and saves it. No shortcuts, no automation scripts, no manual steps.

Dispatch. Claude Dispatch is a remote control feature that connects the Claude mobile app to the Claude Desktop app. You send a task from your phone, and Claude executes it on your desktop computer. The desktop does the actual work: opening files, running commands, interacting with applications. Your phone just provides the instructions and receives the results.

Dispatch is not a remote desktop. You do not see your desktop screen on your phone. You send a text instruction, and Claude sends back the results. It is asynchronous task delegation. You can check on progress from your phone, ask follow up questions, or send additional instructions to the same session.

Computer Use. Claude can control your mouse and keyboard to interact with software just like a human would. It can open your browser, navigate to a website, fill out a form, and submit it. It can open your email client, compose a message, and send it. It can open your spreadsheet, enter data, and save the file.

Computer Use works through screenshots and coordinate based clicking. Claude sees what is on your screen, decides what to click, and simulates the mouse movement and click. It is slower than a human for simple tasks but much faster for repetitive data entry or multi step workflows that require switching between applications.

The current limitations: Computer Use can struggle with applications that have unusual interfaces, custom UI components, or pages that render differently on different screen sizes. It works best with standard web applications and common desktop software.

Clips. Clips are reusable automation templates built on top of Cowork and Dispatch. You create a Clip once by demonstrating the workflow. The Clip saves the sequence of steps. You can trigger it on demand or on a schedule.

Common Clip examples: "Every morning, open my inbox, find emails from my direct reports, summarize any requests, and save the summary to my Desktop." "Before each client meeting, open the CRM, find the account history, open the latest proposal, and compile a briefing document." "Every Friday at 4 PM, check my task list for overdue items, open the related files, and report on status."

Priority access. New features arrive first for Max subscribers. During high traffic periods, Max users get priority compute. For time sensitive work, this matters.

Desktop Apps: Beyond the Web Chat Interface

Both ChatGPT and Claude have desktop applications that go beyond the web interface. This is where the real productivity gains live.

ChatGPT Desktop App

The ChatGPT desktop app (available for macOS and Windows) provides a persistent chat window that stays on top of your work. Key features include:

Voice mode. Hands free conversation with the model. You can dictate prompts and hear responses spoken aloud. Useful when cooking, driving, or doing tasks that occupy your hands.

Screen capture. Take a screenshot of anything on your screen and send it directly to the chat. ChatGPT analyzes the image and responds. Use this for: "Explain this error message," "Turn this chart into a table," "Read this article and summarize it."

App integration. The desktop app integrates with other applications on your computer. You can select text in any app, press a shortcut, and have ChatGPT process it. This works with browsers, email clients, text editors, and most other software.

File access. Open local files from your computer directly in the chat. No need to upload through a web interface.

Claude Desktop App

The Claude desktop app goes further than ChatGPT's because of the Cowork and Dispatch features. Key features include:

Cowork mode. Claude sees your screen and can interact with your applications. You can say "Open the spreadsheet in my Downloads folder, find the row where revenue dropped more than 10%, and explain what the data shows about that month." Claude opens the file, reads it, finds the relevant data, and gives you an analysis.

Clips. Clips are reusable automation templates. You record a workflow once as a Clip, and Claude replays it later. Common Clips include: "Summarize my morning emails," "Create a meeting brief from the last three Slack messages and the calendar event," "Find all unpaid invoices from last month and compile them into a report." Once created, you can trigger a Clip from the mobile app via Dispatch.

Dispatch remote control. This is Claude's signature feature in 2026. Here is how it works in practice. You are on the train commuting to work. You open the Claude mobile app and type: "For my 9 AM meeting, find the proposal draft in my Google Drive, check if the pricing section was updated, and put the latest version on my desktop so I can review it when I arrive." Dispatch sends this task to your desktop at home or office. Claude Cowork opens Google Drive, finds the file, checks the version history, and places the document on your desktop. When you arrive, everything is ready.

Dispatch works for longer running tasks too. Start a research project from your phone while waiting in line. Dispatch triggers Claude to start the work on your desktop. By the time you sit down, the research is done and the results are waiting.

The always on laptop requirement. Dispatch requires your desktop to be awake, unlocked, and running the Claude Desktop app. If your computer goes to sleep, Dispatch fails. Users typically set their computer to never sleep when they plan to use Dispatch remotely. This is a practical consideration. Closing your laptop at the end of the day means Dispatch stops working. Using a desktop Mac mini or keeping a work laptop powered on solves this, but it adds a power and security consideration.

Practical Daily Productivity Examples

Here are specific workflows using the paid features of ChatGPT and Claude.

Email Management (Claude Max Using Dispatch)

Set up a recurring Cowork task that runs every morning at 7 AM. Claude opens your email, reads the last 24 hours of messages, and summarizes them into a briefing document on your desktop. While you commute, check the summary on your phone. Dispatch any follow ups: "Reply to Sarah confirming the meeting time" or "Save the attached contract to the Projects folder."

The time savings compound. A single email triage session that used to take 15 minutes now takes 2 minutes. Over a month, that is over four hours saved.

Meeting Preparation (ChatGPT Pro)

Before a meeting, gather the relevant materials: the calendar invite, the previous meeting notes, any documents shared in the thread. Drag them all into the ChatGPT desktop app. Ask: "Summarize the key decisions from our last meeting. List the action items that were due today. Based on this agenda, what are the three most important topics we need to discuss?"

The chat persists, so you can follow up during the meeting: "Take notes on this discussion and compare them to the agenda." After the meeting: "Draft the meeting summary email based on our discussion."

Document Comparison (Claude Pro for the 200K Context)

You have three vendor proposals, each 40 pages. Upload all three to a single Claude Pro chat. Ask: "Compare these three proposals on pricing, delivery timeline, warranty terms, and cancellation policy. Create a comparison table ranked by total cost of ownership over three years."

The 200K token context handles all three documents in a single conversation. You can ask follow ups about specific sections: "In the second proposal, what is the penalty for late delivery?" "Which proposal has the most favorable payment terms?"

Research and Content Drafting (Either Service, Both Work)

Use Claude Pro's Projects or ChatGPT's Custom GPTs to set up a dedicated workspace for a writing project. Upload your research documents, outline, and style guide. Each time you sit down to write, the project context is already loaded. You do not need to re explain your requirements.

For long form content, use Claude's Projects with custom instructions. Set the instructions to match your writing style: "Write in short paragraphs. Use active voice. Start with a concrete example. Avoid listing features without explaining the benefit." The project remembers these instructions across sessions.

Personal Knowledge Management (ChatGPT Memory)

Use ChatGPT's Memory feature to build a persistent knowledge base about your work and life. Save facts once and reference them later. Tell the model your preferred communication style, the names of your team members, the projects you are working on, and your weekly schedule.

Over time, ChatGPT builds a profile that makes its responses more relevant to your specific situation. When you ask for meeting agenda suggestions, it knows who attends your meetings and what topics are active. When you ask for help drafting an email, it knows the recipient's context from previous conversations.

The tradeoff is privacy. The memory persists across sessions, which means you are trusting OpenAI with personal and professional information. If that makes you uncomfortable, disable the memory feature and use a manual approach instead.

Expense and Receipt Processing (Claude Max Using Computer Use)

Take a photo of a stack of receipts. Send it to Claude via Dispatch. Claude's desktop agent opens your expense tracking spreadsheet or app, reads each receipt, categorizes the expenses, and enters the data. You review the results on your phone and approve the entries.

This workflow turns a tedious 20 minute weekly task into a 30 second photo and approval step. The accuracy depends on receipt quality and handwriting legibility. Blurry or folded receipts may need manual correction.

Data Extraction from Screenshots (ChatGPT Desktop)

Take a screenshot of a table in a PDF that does not allow copy paste. Drag the screenshot into the ChatGPT desktop app. Ask: "Convert this table into a CSV format I can paste into Excel." ChatGPT reads the image, extracts the data, and formats it correctly. This works for printed tables, screenshots of dashboards, and images of documents.

Schedule Management (Claude Pro + Calendar Integration)

Use Claude's integration with your calendar to plan your day. "What does my calendar look like this week? Identify the three most important meetings and suggest preparation steps for each. Flag any scheduling conflicts." Claude reads your calendar, analyzes the events, and gives you a structured overview.

For recurring meetings, set up a Project that contains the meeting context, attendee list, and agenda template. Before each meeting, ask Claude to review the previous meeting's notes from the project and draft the agenda for the current one.

Code Assistance (Claude Code or ChatGPT)

Both services handle code well, but they approach it differently. ChatGPT's code analysis works through the chat interface. You paste code and ask for changes. Claude Code runs in your terminal and can make changes directly to your files.

For non developers, Claude Code is less relevant. But ChatGPT's code features still help: ask it to write a formula for your spreadsheet, create a simple script to rename files, or explain what a line of code in a document does.

Voice to Document (Claude Max)

Use Claude's voice mode to dictate a rough draft of a document while walking or commuting. Dispatch sends the audio transcript to your desktop, where Claude Cowork formats it into a proper document with sections, headers, and formatting. When you get to your desk, the draft is ready to edit.

This workflow is particularly good for first drafts. Dictating removes the friction of staring at a blank page. The content will be rough, but rough content is much easier to edit than to create from nothing.

Which One Should You Choose

The honest answer depends on what you need.

Choose ChatGPT Plus ($20) if you want: The broadest set of features at the standard price point, including image generation, web browsing, and the GPT Store. ChatGPT's ecosystem is larger than Claude's, with more third party integrations and community created tools. If you want one service that does a bit of everything, ChatGPT is the safe pick.

Choose Claude Pro ($20) if you want: Better long form writing, the 200K context window for working with large documents, and the Projects organizational system. Claude tends to produce more natural prose with less prompting, which matters if you do a lot of writing.

Choose ChatGPT Pro ($200) if: You depend on ChatGPT for daily work and consistently hit the Plus limits. You use the API heavily. You want priority access to everything.

Choose Claude Max ($200) if: You want Dispatch, Cowork, and Computer Use. These features are unique to Claude and fundamentally change how you interact with AI. The ability to send tasks from your phone and have them executed on your desktop is a genuine productivity multiplier that no other service offers at this level. If you work across multiple devices and want AI that follows you between them, Claude Max is the clear winner.

Get both if: You have a budget for it and the tasks justify the cost. Many power users maintain subscriptions to both. They use ChatGPT for quick queries, image generation, and web research. They use Claude for writing, document analysis, and remote tasks through Dispatch. The combined $40/month (or $400 if you go Max and Pro) covers almost every AI use case.

But the honest truth is that most people do not need to pay $200 a month for either service. Start with one $20 subscription. Use it for a month. If you hit limits or want features you do not have, consider upgrading or adding the second service.

What Comes Next

Part 4 of this series moves beyond text assistants into specialized AI tools. We will cover music generation with Suno and Udio, video creation with Veo and Runway, image generation with Midjourney and DALL E, and how these creative tools can fit into your daily productivity workflow.

Continue to Part 4: Specialized AI Tools for Creation

Skip to Part 5: Going Advanced: Open Source, Local Models, and Agent Tools

A Tour of Specialized AI Tools: Music, Video, Images, and More

Alex Merced — Mon, 01 Jun 2026 09:00:00 GMT

The first three parts of this series covered general purpose AI assistants: the chatbots and writing tools that handle text based tasks. But AI in 2026 extends far beyond chat windows. A whole ecosystem of specialized tools creates original music, generates cinematic video, produces professional images, and designs presentations.

This is Part 4 of "Catching Up with Using AI for All Levels." If you are new here, start with Part 1 for the fundamentals and Part 2 for the free tools, then Part 3 for ChatGPT and Claude. This post covers the creative side: what the tools are, what they cost, how good the output actually is, and when they make sense for daily productivity rather than just artistic projects.

Skip to Part 5: Going Advanced: Open Source, Local Models, and Agent Tools

Image Generation: The Mature Category

Image generation is the most mature of the creative AI categories. It started with DALL E 2 in 2022 and has grown into a competitive market with multiple strong options at various price points. The quality has improved to the point where AI generated images are used professionally in marketing, publishing, and product design.

Midjourney

Midjourney remains the gold standard for artistic quality. It operates through Discord, which is both its strength and its biggest friction point. You join the Midjourney Discord server, type your prompt in a channel, and the bot generates images in the thread.

The Discord interface has improved over time. The bot now supports private messaging, so you do not need to share your generations with strangers. The web gallery is well organized. But the overall experience still feels like a workaround for the lack of a native application.

Midjourney's output quality is excellent. The model understands composition, lighting, color theory, and artistic style better than any competitor. Its strength is producing images that look like professional photography or illustration work. If you need a photorealistic product shot in a specific lighting setup, Midjourney delivers. Its weakness is that it struggles with precise text rendering and specific brand requirements. Do not ask Midjourney to generate an image with a specific word or logo displayed clearly. It will get close but not exact.

Pricing starts at $10 per month for Basic (3 hours of GPU time, roughly 200 images) and goes up to $60 per month for Mega (60 hours of GPU time). The Standard plan at $30 per month is the sweet spot for regular users.

DALL E 4 (via ChatGPT)

DALL E 4 is OpenAI's latest image generation model, available through ChatGPT Plus ($20/month). It has improved significantly over DALL E 3, with better prompt adherence, more consistent anatomy, and improved text rendering. DALL E 4 can render short words and phrases legibly in images, a longstanding weakness of earlier AI image generators.

DALL E 4's main advantage is integration. Because it lives inside ChatGPT, you can iterate naturally. Generate an image, discuss it with the model, request changes, and generate the next version, all in one conversation. This tight feedback loop makes it the most efficient image generator for most workflows, even if Midjourney produces better standalone results.

For productivity, the ChatGPT integration is the killer feature. You can be drafting a presentation slide in ChatGPT, generate an accompanying image in the same conversation, and get suggestions for how to arrange both on the slide. No switching between tools. No copying prompts between applications.

Flux and Stability AI

Flux, created by Black Forest Labs (a team of former Stability AI researchers), has emerged as a strong open weights competitor. Flux Pro is available through Fireworks AI and other providers. It competes with Midjourney on quality while being accessible through developer friendly APIs. Flux is available in several variants: Flux Pro for highest quality, Flux Dev for faster generation, and Flux Schnell for rapid prototyping.

The open weights nature of Flux means you can run it on your own hardware or through any provider that hosts it. This flexibility makes it popular with developers who want to integrate image generation into their own applications without per image API fees from a single vendor.

Stability AI continues to develop Stable Diffusion, the open source image generation model. Stable Diffusion 4 is available in 2026 with strong quality and the advantage of running locally on consumer GPUs. For users who want privacy, offline access, and no subscription fees, Stable Diffusion remains the best option. The tradeoff is that running it locally requires a capable GPU and some technical setup.

Nano Banana

Nano Banana has gained attention in 2026 as a new contender. It produces high quality images with a simple interface and competitive pricing. The Pro version includes upscaling, inpainting, and style transfer. It is worth trying alongside Midjourney and DALL E to see which style fits your needs. Nano Banana's strength is its ease of use for non technical users who want good results without learning complex prompt engineering.

Who Should Use Image Generation

The most practical productivity use is creating visuals for presentations, social media, and internal documents. Instead of spending 30 minutes searching stock photo sites for the right image, you generate exactly what you need in 30 seconds.

Business use cases include: product mockups for proposals, custom illustrations for blog posts, branded social media graphics, concept visualizations for client presentations, and placeholder images for website designs.

The quality is good enough for professional use in most contexts, but you should still use real photography for anything that represents an actual product, person, or location. AI generated images have subtle tells that trained eyes notice.

Video Generation: The Fastest Evolving Category

Video generation has progressed faster than any other AI category in the last 18 months. We went from short, glitchy clips to multi minute videos with consistent characters, coherent motion, and usable quality.

Veo 3.1 (Google)

Veo 3.1 is widely considered the best overall AI video generator in 2026. It produces high resolution videos with strong prompt adherence, consistent character appearance across frames, and minimal artifacts. The improvement from Veo 2 to Veo 3.1 was dramatic, with much better motion coherence and fewer of the warping artifacts that plagued earlier video models.

Veo is available through Google AI Studio with a free tier that lets you generate short clips for testing. Full access requires a Google AI Premium ($20/month) or AI Ultra ($100/month) subscription. The free tier is generous enough for experimentation and light use, making Veo the most accessible high quality video generator.

Veo excels at text to video: describe a scene and get a video clip. It also supports image to video: upload an image and animate it. The image to video feature is particularly useful for creating short animations from still photographs or illustrations. You can take a product photo and generate a slow orbit around it, turning a static image into a dynamic product showcase.

Veo's main limitation is creative control. You describe what you want and accept what you get. There is no way to fine tune the motion, adjust camera angles, or edit specific frames. For one shot generation where speed matters more than precision, Veo is the best choice.

For productivity, Veo is useful for creating short explainer videos, social media content, and presentation clips. A 15 second product demo video that used to take a day to produce can now be generated in minutes. The quality is good enough for social media and internal use, though not yet at broadcast quality.

Runway Gen 4

Runway is the veteran of AI video generation, having launched Gen 1 in early 2023. Gen 4, released in late 2025, offers the most creative control of any video generator. It includes features like Motion Brush (paint movement onto specific areas of an image), Act One (transfer facial expressions from a reference video), and inpainting (edit specific areas of a generated video).

The Motion Brush is Runway's standout feature. You upload an image, paint a brush stroke across the area you want to animate, and define the direction and speed of movement. Want smoke rising from a chimney in a still photo? Paint the chimney area and set the motion upward. Want water flowing in a river? Paint the river surface and set the flow direction. This level of granular control is unique to Runway.

Runway's pricing starts at $15 per month for the Standard plan with limited credits (enough for roughly 50 short generations). The Pro plan at $35 per month gives more credits and higher resolution output. For heavy use, the Unlimited plan at $95 per month removes credit limits. The pricing is higher than Veo's bundled cost, but the additional control justifies the premium for professional use.

Runway is the best choice when you need precise control over the output. If you need a specific camera movement, a particular character action, or an edit to an existing generation, Runway gives you the tools to iterate. Veo is better for one shot generation where you describe what you want and accept the result.

Kling AI

Kling, developed by Kuaishou (the company behind a major Chinese video platform), has emerged as a strong competitor. It offers high quality video generation at competitive prices, with particularly good results for character animation and cinematic shots.

Kling uses a credit system with free trial credits and paid packs starting around $10. The quality is comparable to Veo and Runway for many use cases, though it lags slightly on text rendering and complex scene composition.

Who Should Use Video Generation

Video generation is still more of a content creation tool than a daily productivity tool for most people. The practical use cases are concentrated in marketing, content creation, education, and internal communication.

A non obvious productivity use: creating quick tutorial videos for your team. Instead of writing a 3 page document explaining how to use a new process, generate a 60 second video walkthrough. The video is easier to consume and more likely to be watched than a document to be read.

The current limitations are real. Videos longer than 30 seconds still struggle with consistency. Characters in the first frame may change appearance by the tenth frame. Complex action sequences produce artifacts. Text rendering in video is unreliable. You should budget time for multiple attempts and manual editing to get a usable result.

Music Generation: From Novelty to Useful

AI music generation has come into its own in 2026. The tools produce genuinely listenable songs with vocals, multiple instruments, and coherent structure.

Suno

Suno is the leading AI music generator. It generates complete songs with lyrics, vocals, and instrumentation from a text prompt. You describe the genre, mood, and subject, and Suno produces a full track with verses, choruses, and a bridge.

Suno's free tier gives you a limited number of generations per day, enough for experimentation. The Pro plan at $10 per month gives 500 generations and commercial usage rights. The Premier plan at $30 per month gives 2,000 generations and priority processing.

The output quality varies by genre. Pop, rock, electronic, and hip hop work well. Classical and jazz are less convincing. The vocals sound synthetic on close listening but pass for casual listening in the background. The instrumental quality is generally strong.

The most practical productivity use for Suno is creating custom background music for videos, presentations, and internal content. Instead of searching royalty free music libraries for the right track, you generate a track that matches your specific needs: "Upbeat electronic background music, 120 BPM, no vocals, suitable for a tech product demo."

Udio

Udio is Suno's primary competitor. It produces comparable quality with a slightly different emphasis. Udio excels at creative exploration: you can remix existing songs, extend specific sections, and edit the structure more granularly than Suno allows.

Udio's pricing is similar to Suno's, with a free tier and paid plans starting at $10 per month. The choice between Suno and Udio comes down to personal preference for the output style and the editing workflow you prefer.

AIVA

AIVA specializes in orchestral and cinematic music. If you need a string quartet arrangement, a film score style piece, or ambient orchestral background music, AIVA produces the most convincing results in this niche.

AIVA has a free tier for limited generations. Paid plans start at $15 per month for higher quality output and commercial rights. It is less versatile than Suno or Udio but better within its niche.

Who Should Use Music Generation

Music generation is the most situational of the creative AI tools. If you create any kind of video content, presentations, podcasts, or social media posts, generating custom background music saves time and avoids copyright issues.

The hidden productivity use is inspiration and mood setting. Generate a few short musical pieces for a creative project and use them as background while you work. The music sets a tone that helps you get into the right mental state for the task.

Presentation and Design Tools

Beyond the obvious image, video, and music generators, a category of AI tools focuses specifically on business productivity tasks.

Gamma

Gamma creates presentations, documents, and web pages from a single prompt. You describe what you need, and Gamma generates a complete deck with text, images, and layout. The output is good enough for internal presentations and early stage client work.

Gamma's free tier allows a limited number of generations. The Pro plan at $16 per month removes most limits and adds higher resolution exports.

The productivity gain is significant for anyone who regularly creates presentations. A deck that takes two hours to build manually takes five minutes with Gamma. The tradeoff is that the output looks like an AI generated deck: competent but generic. Gamma is best for first drafts that you then customize with your own branding and specific content.

Beautiful AI

Beautiful AI predates the current generative AI wave. It uses AI for layout and design recommendations rather than content generation. You add text and images manually, and the AI arranges them into professional looking slides.

Beautiful AI complements Gamma well. Use Gamma to generate the first draft, then import it into Beautiful AI for layout refinement. The combination covers both content generation and visual polish.

Canva AI

Canva has integrated AI features across its entire platform. Magic Design generates complete designs from text prompts. Magic Eraser removes unwanted objects from images. Magic Expand extends image boundaries. Magic Write generates and edits text.

Canva's AI features are available on the free tier with usage limits. The Pro plan at $13 per month removes most limits and adds brand kits, background removal, and premium templates.

Canva AI is the most practical choice for non designers who need to create visual content regularly. The learning curve is minimal, the output quality is good, and the integrations with social media platforms streamline publishing.

Audio: Transcription, Voice, and Sound

Descript

Descript is an audio and video editor with AI transcription at its core. You upload an audio or video file, Descript transcribes it, and you edit the media by editing the text. Delete a sentence from the transcript and the corresponding audio is removed. Change a word and Descript regenerates the audio in the original speaker's voice.

The workflow is simple. Import a recording of a meeting, podcast, or voiceover. Descript transcribes everything automatically, usually within a few minutes for a one hour recording. You see the full transcript with speaker labels and timestamps. Edit the transcript as you would edit a document: delete filler words, reorder sections, fix mispronounced terms. The audio and video update automatically to match your text edits.

Descript also includes AI voice generation (Studio Sound), noise reduction, and filler word removal. The Studio Sound feature analyzes your recording and removes background noise, echo, and room tone. It is good enough to make a recording from a noisy coffee shop sound like it was recorded in a treated studio.

The free tier covers basic transcription with limited exports. The Pro plan at $24 per month adds screen recording, unlimited transcription, and Studio Sound. The Business plan at $40 per month adds team features and brand voices.

For productivity, Descript is invaluable for anyone who creates audio or video content. Editing a spoken recording by editing text is dramatically faster than working with audio waveforms. The filler word removal alone saves 20 minutes per hour of recording. For meeting recordings, Descript generates searchable transcripts that let you find any topic discussed in a one hour meeting within seconds.

ElevenLabs

ElevenLabs is the leading AI voice generation platform. It produces the most natural sounding synthetic voices available, with accurate emotion, pacing, and emphasis. The voice cloning feature lets you create a digital copy of your own voice from a short recording, as little as 30 seconds of audio.

The quality has improved to the point where short AI generated voice clips are difficult to distinguish from human recordings. Longer passages still have subtle tells: slightly unnatural pacing, odd emphasis on certain words, and a lack of breath sounds at natural intervals. But for most practical purposes, the quality is sufficient.

ElevenLabs pricing starts at $5 per month for the Starter plan with limited characters (roughly 30 minutes of generated speech). The Creator plan at $22 per month is suitable for regular use with longer character limits. The Pro plan at $99 per month is for high volume commercial use.

Productivity use cases include: generating voiceovers for videos and presentations, creating audio versions of written content (your blog posts, newsletters, internal memos), adding narration to tutorials and training materials, and producing multilingual versions of existing audio content. ElevenLabs supports 29 languages with good quality across most of them.

A Note on Ethics

Voice cloning raises obvious ethical concerns. You should only clone a voice with the person's explicit consent. Using ElevenLabs to impersonate someone without permission is not just unethical. It could be illegal in some jurisdictions. The platform has safety measures in place, including voice authentication and content moderation, but the responsibility ultimately rests with the user.

Adobe Firefly and the Creative Suite Integration

Adobe has integrated AI into its Creative Cloud suite through Firefly, its generative AI engine. Photoshop includes Generative Fill and Generative Expand, which let you add or remove elements from an image with text prompts. Illustrator has Generative Recolor and text to vector graphics. Premiere Pro includes text based editing similar to Descript.

Firefly is notable because it is trained on Adobe Stock images and openly licensed content, which means the output is cleared for commercial use. If you work in marketing, publishing, or any context where copyright ownership matters, Firefly's training data provenance gives it an advantage over models trained on scraped internet data.

Firefly is included in existing Creative Cloud subscriptions. Photoshop users with a subscription get a certain number of generative credits per month. Additional credits are available for purchase.

Putting It All Together: A Creative AI Workflow

Here is how these tools work together for a real project.

You need to create a product launch video for a new software feature. Start with Suno to generate background music: "Upbeat electronic track, 90 BPM, no vocals, 60 seconds with a clear crescendo at the 45 second mark." Download the track.

Use Midjourney to generate key visual frames: "A person using a laptop with a glowing screen, clean modern office, cinematic lighting, photorealistic." Select the best images.

Upload the images to Veo or Runway. Generate short animated clips from each image: "Camera slowly zooming in on the screen." Combine the clips.

Use ElevenLabs to generate a voiceover from your script. Import the voiceover, music, and video clips into Descript. Edit by editing the transcript. Fine tune the timing. Export the final video.

The entire workflow takes two to three hours for a 60 second launch video. The same project with traditional tools would take a full day or more, depending on your skill level with each medium.

Internal Training Videos

Your company needs a short training video explaining a new expense reporting process. Start with Gamma to generate a presentation deck with the key steps. Use ElevenLabs to generate a voiceover from the deck text. Use Suno to generate background music. Use Runway's image to video feature to animate any static diagrams. Combine everything in Descript for final editing. Total time: two hours for a five minute training video that would have taken a day and a half with traditional tools.

Social Media Content Calendar

You manage social media for a small business. Each week you need 5 images, 5 captions, and maybe one short video. Use Midjourney or DALL E to generate consistent branded images. Set a style reference in your prompts to keep visual consistency across posts. Use ChatGPT to draft captions in your brand voice. Use Veo to generate one 10 second video clip per week showcasing a product or service. Use Canva AI to arrange everything into the correct dimensions for each platform. The weekly content that used to take 4 hours now takes 45 minutes.

Client Proposal with Visuals

You are preparing a client proposal. Write the content in ChatGPT or Claude. Generate relevant diagrams and concept images in DALL E or Midjourney. If the proposal involves a physical product, generate a short Veo animation showing the product from multiple angles. Combine everything in Canva or Gamma for the final presentation. The result is a professional, visually rich proposal that looks like it took days to produce, completed in a few hours.

Personal Photo Projects

For personal use, the free tiers of these tools cover most needs. Edit family photos with Photoshop's Generative Fill to remove photobombers or improve composition. Use Google Photos AI to organize and search your library. Use Suno's free tier to generate a custom song for a friend's birthday. Use Canva AI to design invitations, cards, and social media posts for personal events.

The Real Productivity Question

The question to ask about any specialized AI tool is not "Can it generate what I need?" It is "Does it save me more time than it costs?"

The cost is not just the subscription price. It is the time spent learning the tool, the time spent iterating on prompts to get the output you want, and the time spent fixing problems that the AI introduced.

For a business user who creates presentations and social media graphics regularly, Canva AI and Gamma are clear wins. The time saved per task is dramatic, and the learning curve is shallow.

For a content creator who publishes weekly videos, Descript and Runway are worth the investment. The production speed increase pays for the subscriptions many times over.

For someone who generates music or images recreationally, the free tiers are generous enough that you can explore without commitment. Pay only when you hit the free limits and find yourself wishing for more.

Part 5 of this series covers the most advanced territory: open source models that run on your own computer, agent frameworks like Hermes Agent, and coding tools like OpenCode. These tools require more setup but offer privacy, offline access, and capabilities that cloud services cannot match.

Continue to Part 5: Going Advanced: Open Source, Local Models, and Agent Tools

Going Advanced: Open Source Models, Hermes Agent, and Local AI

Alex Merced — Mon, 01 Jun 2026 09:00:00 GMT

This is the final installment of "Catching Up with Using AI for All Levels." Parts 1 through 4 covered the fundamentals, free tools, paid services, and specialized creative tools. This post goes deeper. We will explore the open source ecosystem: models you can download and run on your own computer, agent frameworks that automate complex tasks, and coding tools that work entirely offline.

This is the most technical post in the series, but do not let that scare you. The tools have matured significantly in 2026. Installing and running a local AI model is easier than it was six months ago, and the benefits are real: privacy, offline access, no subscription fees, and unlimited usage after the initial hardware investment.

Return to Part 4: Specialized AI Tools for Creation

Why Run AI on Your Own Hardware

Before we get into the tools, it is worth understanding why someone would choose local AI over the convenience of cloud services.

Privacy. When you use ChatGPT, Claude, or Gemini, your conversations are processed on someone else's servers. The companies store and analyze your data. For personal use, this might be acceptable. For business use, especially with sensitive or proprietary information, it is a deal breaker. Local models process everything on your computer. Nothing leaves your machine.

Offline access. Cloud services require an internet connection. Local models work anywhere: on a plane, in a remote area, in a secure facility with no external network access. If connectivity is unreliable where you live or work, local AI is the only option.

No subscription fees. The cloud services cost $20 or $200 per month. Local models cost nothing to use after you buy the hardware. If you do the math over three years, a $2,000 computer running local models is cheaper than $720 of ChatGPT Plus or $7,200 of Claude Max.

Unlimited usage. Cloud subscriptions have rate limits. Pro users hit them regularly. Local models have no rate limits. You can use them as much as you want, as fast as your hardware allows.

Customization. Open source models can be fine tuned, modified, and specialized for your specific needs. You can train a model on your own documents. You can adjust its behavior in ways that cloud APIs do not permit.

The tradeoffs are performance and capability. Local models are slower than cloud models. They are less capable, especially at complex reasoning tasks. A 7 billion parameter model running on a laptop cannot match GPT 5.4 or Claude Opus. But the gap has narrowed significantly, and for many everyday tasks, local models are good enough.

The Hardware You Need

Running local AI requires a capable computer. The most important component is the GPU (graphics processing unit), because AI inference runs much faster on GPUs than on CPUs.

Minimum setup: 8GB of RAM and a modern CPU. You can run small models (1 to 3 billion parameters) on CPU alone. They will be slow, taking 10 to 30 seconds per response, but they work. This is enough for basic text generation and summarization. Models like Phi 4 (14B) in 4 bit quantization can run on CPU with acceptable speed for occasional use.

Good setup: 16GB of RAM and a GPU with 8GB of VRAM, such as an RTX 3060 or RTX 4060. This handles 7 to 8 billion parameter models comfortably. Response times are 2 to 5 seconds. Most mid range gaming laptops and desktop GPUs from the last few years meet this requirement. This is the sweet spot for most users: you can run Llama 4 8B, Qwen 3.5 7B, or DeepSeek R1 7B with good performance.

Great setup: 32GB of RAM and a GPU with 16 to 24GB of VRAM, such as an RTX 4090, RTX 5090, or a used RTX 3090. This handles 13 to 30 billion parameter models. Response times are under 2 seconds. Models like DeepSeek R1 14B, Qwen 3.5 32B, and Mistral Small 4 24B run well here.

Overkill: 64GB+ RAM and 48GB+ VRAM, such as dual RTX 5090s or professional GPUs like the NVIDIA A series. This handles 70+ billion parameter models, including the largest open weight models like Llama 4 405B or DeepSeek V4 671B (mixture of experts, so only part of the model activates at a time). Response times are comparable to cloud services for many tasks.

The good news is that model quantization has improved dramatically. Quantization compresses models to use less memory with minimal quality loss. A 70 billion parameter model that used to require 140GB of VRAM with full precision can now run on 24GB with 4 bit quantization while retaining most of its capability.

What You Can Expect at Each Level

At the minimum level, you get a capable assistant for simple tasks. It can summarize short documents, draft emails, answer questions about general knowledge, and help with basic coding. The responses are slower but usable.

At the good level, you get a solid daily driver. It handles most tasks well: document analysis, longer writing, moderate reasoning, and code generation. The speed is good enough for interactive use without frustration.

At the great level, you approach cloud model quality for many tasks. Complex reasoning, multi step instructions, and long context windows work well. The difference between this and GPT 5.4 or Claude Opus is noticeable on hard tasks but acceptable for everyday use.

CPU Only: Is It Worth Trying?

If you do not have a GPU, you can still run local models on CPU. The experience depends on the model size and your patience. Small models (1 to 3B parameters) run at reasonable speed. Medium models (7 to 8B) run at 3 to 10 tokens per second on a modern CPU, which means 10 to 30 seconds per response. This is usable for batch processing or tasks where speed does not matter, but it is too slow for interactive conversation.

The CPU path is worth trying to understand the ecosystem before investing in a GPU. Install Ollama, download a small model, and see what local AI feels like.

Ollama: The Easiest Way to Run Local Models

Ollama is the simplest tool for running local LLMs. It handles model downloading, management, and inference through a single command line interface. It also serves an OpenAI compatible API, which means any application that works with OpenAI's API can work with your local models.

Installation

Ollama runs on macOS, Linux, and Windows. Download the installer from ollama.com and run it. The installation takes about two minutes. After installation, Ollama runs as a background service.

Running a Model

Open a terminal and run:

ollama run llama4

Ollama downloads the model (a few gigabytes) and starts an interactive chat session. Type your questions and get responses from the local model. Type /exit to leave.

The first download takes time depending on your internet speed. Subsequent runs use the cached model and start instantly.

Available Models

Ollama hosts hundreds of models. Here are the ones worth knowing about in 2026.

Llama 4 (Meta). Meta's latest open model family ranges from 8 billion to 405 billion parameters. The 8B model runs on modest hardware and handles general conversation, summarization, and simple tasks well. The 70B and 405B models require powerful hardware but approach cloud model quality.

DeepSeek V4 and R1. DeepSeek's models have been a sensation in the open source community. V4 is their general purpose model, comparable to GPT 4 class models. R1 is their reasoning model that shows its chain of thought before answering. Both are available in sizes from 7B to 671B (mixture of experts). The smaller distilled versions (7B and 14B) run on consumer hardware and punch above their weight class.

Qwen 3.5 and 3.6 (Alibaba). Qwen models have consistently improved. The 3.5 series offers strong performance across coding, reasoning, and general tasks. Qwen 3.6, released in mid 2026, adds improved instruction following and longer context handling. The 7B and 14B versions are popular choices for local deployment.

Gemma 4 (Google). Google's open model family is designed for efficient inference. Gemma 4 26B performs well for its size and runs on a single 24GB GPU. The smaller Gemma 4 9B runs on 8GB hardware.

Mistral Small 4 and Mixtral. Mistral's models are known for efficiency. Mistral Small 4 (24B) punches above its weight. The Mixtral 8x22B mixture of experts model offers strong performance with efficient resource usage.

Phi 4 (Microsoft). Microsoft's Phi series focuses on small, capable models. Phi 4 runs on modest hardware and handles coding and reasoning tasks surprisingly well for its size.

Practical Example

Install Ollama, download Llama 4 8B, and start using it as a local assistant. Ask it to summarize documents (paste the text directly), draft emails, explain concepts, or brainstorm ideas. The quality is not as good as GPT 5.4, but it is fast, private, and free.

For better quality, try DeepSeek R1 14B. The chain of thought reasoning makes it more thorough for complex questions, and it runs well on 12GB of VRAM.

LM Studio: A Graphical Interface for Local Models

If the command line is not your preference, LM Studio provides a graphical interface for running local models. It includes a model browser, download manager, and chat interface in a single desktop application.

LM Studio also serves an OpenAI compatible API endpoint, just like Ollama. You can start the local server and point any application at http://localhost:1234/v1. This is useful for connecting local models to other tools.

The key advantage of LM Studio is LM Link, a feature that lets you access a model running on one computer from other devices on your network. You can run a large model on your powerful desktop and access it from your laptop. This uses Tailscale for secure tunneling and works across networks.

DeepSeek: The Open Weight Powerhouse

DeepSeek deserves special attention because it changed the open source AI landscape. In early 2025, DeepSeek released R1, a reasoning model that matched OpenAI's best models at a fraction of the training cost. The company followed with V3 and V4, each improving on the last.

DeepSeek models are open weight, meaning you can download the actual trained parameters and run them on your own hardware. This is different from open source models that release only the architecture and training code. Open weight models let you inspect, fine tune, and deploy the exact model that achieved specific benchmark results.

DeepSeek V4 is available through various providers. You can access it through chat.deepseek.com for free with rate limits, through API providers for pay per use pricing, or download it for local deployment.

The pricing advantage is substantial. DeepSeek's API costs roughly 10 to 20 times less than OpenAI's API for comparable quality. This makes it attractive for developers building applications that make many API calls, and for users who want to experiment with advanced models without committing to a $200 per month subscription.

OpenCode: Terminal Based Coding Agent

OpenCode is an open source, provider agnostic coding agent that runs in your terminal. It is similar to Claude Code but works with any model provider, including local models.

Installation

OpenCode installs with a single command. On macOS: brew install opencode. On Linux and Windows, the website provides installers. The setup takes under a minute.

Configuration

OpenCode works out of the box with cloud providers by default. To use it with a local model, create an opencode.jsonc configuration file that points to your local endpoint.

json { "provider": { "local": { "name": "Local Model", "npm": "@ai-sdk/openai-compatible", "options": { "baseURL": "http://localhost:11434/v1" }, "models": { "deepseek-r1-14b": { "name": "DeepSeek R1 14B", "modalities": { "input": ["text"], "output": ["text"] } } } } } }

This configuration tells OpenCode to use a model running on Ollama's local endpoint.

Daily Use

OpenCode operates in two modes. Interactive mode (opencode) opens a terminal UI where you describe your task and OpenCode works through it step by step. One shot mode (opencode run "Add error handling to the API calls in src/client.py") executes a single task and exits.

For developers, OpenCode replaces or supplements GitHub Copilot and Claude Code. It understands your project structure, reads and writes files, runs terminal commands, and manages git operations. The key advantage is provider flexibility: you can use a local model for simple tasks and switch to a cloud model for complex ones, all within the same tool.

For non developers, OpenCode is less directly useful. But it powers other applications that may benefit you indirectly, such as automated document processing pipelines and data transformation tools.

Hermes Agent: The AI Orchestrator

Hermes Agent, created by Nous Research, is an open source AI agent framework that runs in your terminal. Think of it as an AI assistant that can use tools, remember context across sessions, and coordinate with other AI tools.

What Makes Hermes Different

Hermes is provider agnostic. It works with any LLM provider: OpenAI, Anthropic, OpenRouter, or local models via Ollama. You configure the provider once, and Hermes handles the rest.

The persistent memory system is a standout feature. Hermes remembers facts across sessions. Tell it your preferences once, and they apply to all future conversations. This is similar to ChatGPT's memory feature but runs entirely on your machine with no data sent to a cloud service.

The skill system lets you save reusable procedures. If you have a workflow you repeat often, you can save it as a skill, and Hermes runs it automatically when the relevant context appears. Skills are just markdown files that describe the workflow, making them easy to create and modify.

Installation

bash curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup

The setup walks you through configuration: choosing a provider, setting up tools, and configuring your preferred model.

Daily Use

Use Hermes in three ways.

Interactive chat. Run hermes to start an interactive session. You can ask questions, request tasks, and have multi turn conversations, just like ChatGPT or Claude, but with the ability to run commands, access files, and use tools.

One shot queries. hermes chat -q "Summarize the changes in the last three git commits" runs a single query and returns the result. This is useful for scripting and automation.

Scheduled tasks. Hermes can run tasks on a schedule using cron. Set up a daily briefing that summarizes your calendar, email, and task list every morning. The output can be delivered to Slack, Telegram, email, or saved to a file.

Daily Productivity Examples for Non Developers

For readers who are not developers, Hermes might sound like a developer tool. It is, but its uses extend beyond coding. Here are concrete examples that anyone can use.

Research automation. Save a research workflow as a Hermes skill. When you need to research a topic, run the skill, and Hermes searches the web, extracts key information, and compiles a summary. You do not need to manually search, copy, and paste. The skill remembers your preferred format: bullet points, paragraph summaries, or structured reports.

File organization. Ask Hermes to organize your Downloads folder by file type, date, or project. "Find all PDFs modified in the last week and move them to a Research folder. Delete any .tmp files older than 30 days. Create subfolders by category based on file names." Hermes executes the task using its terminal access, moving and organizing files according to your instructions.

Meeting notes. Record meeting notes as a Hermes skill. When you finish a meeting, run the skill and Hermes prompts you for key decisions, action items, and follow ups. It formats the output consistently and saves it to your notes folder with a timestamp and meeting title. Over time, you build a searchable archive of formatted meeting notes.

Content repurposing skill. Save a skill that takes a piece of content and produces versions for different platforms. Input: a blog post. Output: LinkedIn summary, Twitter thread, newsletter excerpt, and internal memo. Run it once, get all four formats. The skill defines the tone and length for each platform so you do not need to repeat instructions.

Weekly review. Set up a scheduled Hermes task that runs every Friday at 4 PM. It reviews your week's activity, summarizes what you accomplished, and drafts a status report. You review and send. The cron based scheduling means the task runs automatically without you remembering to start it.

The key insight is that Hermes remembers. Unlike ChatGPT or Claude, which start fresh each conversation, Hermes saves skills and memories. Every hour you invest in setting up skills pays back in future sessions as tasks that used to take ten minutes now take one.

Privacy First: The Local AI Stack

The most interesting development in 2026 is the local AI stack: a fully private, offline setup that replaces cloud dependent tools. The standard combination is:

LM Studio or Ollama to run models locally. OpenCode for coding tasks, pointing at the local model. Hermes Agent for general purpose tasks and orchestration, also pointing at the local model.

All three tools can use the same local model through Ollama's API. You install Ollama once, download a few models, and both OpenCode and Hermes connect to it automatically.

The result is a fully private AI setup. Your data never leaves your machine. No subscriptions. No rate limits. No privacy concerns. The tradeoff is capability: local models are not as smart as the cloud frontier models. But for many everyday tasks, they are good enough.

The author of a popular blog post about this setup summarized it well: "Hermes has an OpenCode skill, which means it can fire up OpenCode and interact with it." The orchestrator delegates complex tasks to the specialized tool, and the whole system works together.

Open Source Image Generation

Local AI is not limited to text. Image generation models also run on your own hardware.

Stable Diffusion 4 runs locally via AUTOMATIC1111's WebUI, ComfyUI, or InvokeAI. You download the model once and generate unlimited images with no per image costs. The quality is competitive with cloud services for many use cases.

Flux is available in open weights through Black Forest Labs. The smaller Flux Dev model runs on 12GB VRAM and produces high quality images. Flux Schnell is a faster variant for rapid prototyping.

Running image generation locally requires a GPU with sufficient VRAM. 8GB handles Stable Diffusion and Flux Schnell. 16GB handles Flux Dev and higher resolution outputs.

The Bottom Line

The open source AI ecosystem in 2026 is mature enough for daily use. You can run a capable language model, a coding assistant, an agent framework, and an image generator all on a single consumer grade computer, all for free after the hardware purchase.

For anyone who values privacy or needs offline access. Anyone who wants unlimited usage without subscription fees. Anyone who enjoys tinkering with technology and wants full control over their AI tools.

The privacy argument is the strongest for most users. Consider this: every prompt you type into ChatGPT becomes part of OpenAI's training data unless you opt out. Every document you upload to Claude is processed on Anthropic's servers. For personal use, you might be comfortable with this. For work related tasks, your employer's policies may prohibit sending data to third party AI services. Local models remove this concern entirely.

Who should stick with cloud services? Anyone who wants the best possible quality with zero setup effort. Cloud models are still smarter, faster, and more reliable than local alternatives. For mission critical work where quality matters most, the cloud is the better choice. If you are a writer producing polished content, a researcher synthesizing complex information, or a developer solving hard problems, the $20 per month for ChatGPT Plus or Claude Pro is money well spent.

Who should try local? Anyone who is curious, values privacy, works offline sometimes, or wants to avoid ongoing subscription costs. The setup takes an afternoon. The hardware costs money upfront but pays for itself over time. And the learning process itself is valuable: running a local model gives you a deeper understanding of how AI actually works, which is the whole point of this series.

The best approach is hybrid. Use cloud services for the hard stuff: complex reasoning, long form writing, creative brainstorming. Use local models for everyday tasks: simple questions, document summaries, quick drafts, and anything involving sensitive data. Both have their place, and the tools now make it easy to switch between them.

Quick Decision Guide

Goal	Recommendation
Best quality, no setup	ChatGPT Plus ($20/mo) or Claude Pro ($20/mo)
Best quality, remote desktop control	Claude Max ($200/mo) for Dispatch and Cowork
Privacy, offline, zero recurring cost	Ollama + local models, free after hardware
Best balance of privacy and capability	Hybrid: cloud for hard tasks, local for sensitive work
Developer coding assistant	OpenCode or Claude Code with your choice of model
Full AI automation on your machine	Hermes Agent with skills and scheduled tasks

The Takeaway

You do not need to choose one approach exclusively. The best AI setup in 2026 uses multiple tools for different tasks. Free Google services for everyday queries. A $20 subscription for hard problems. Local models for sensitive work. Specialized tools for creative projects. The ecosystem is broad enough that there is a right tool for every task and a price point for every budget.

Where to Go From Here

This series covered the full spectrum of AI tools available in 2026. You started with the fundamentals in Part 1: how prediction engines, vectors, and transformers work. You moved to free tools in Part 2: Gemini, NotebookLM, and the Google AI ecosystem. Part 3 covered the paid powerhouses: ChatGPT and Claude with their desktop apps, Dispatch, and advanced features. Part 4 toured the creative side: image, video, music, and audio generation. This final part opened the door to the open source world: local models, agent frameworks, and complete privacy.

The point of this series is not to sell you on any particular tool or approach. It is to give you a map of the landscape so you can choose what fits your needs. The best AI tool is the one you actually use. Start with the free tier. Add a paid subscription when you hit limits. Explore local models if privacy matters to you. Switch between tools depending on the task.

AI is not magic and it is not sentient, as we covered in Part 1. It is a tool, like a search engine or a spreadsheet, but more versatile than either. The more you understand what it is and what it is not, the better you will use it.