---
title: "The AI Tool Map: Pick the Right Tool for the Job"
author: "Ecem Karaman"
source: "https://aiwithecem.com/guides/ai-tool-map"
published: 2026-09-16
tools: ["Multi-Tool"]
topics: ["Setup","Workflows","Building","Creative"]
---

# The AI Tool Map

There are more AI tools than anyone can test. This map organizes the useful ones by the job you need done, so you can pick one or two options instead of reading reviews of fifty.

**Last checked against each tool's official site on September 16, 2026.** Plans and features change often, so confirm the price on the tool's own pricing page before you pay. Nothing here is sponsored.

> **In this guide**
>
> 1. [How to use this map](#how-to-use-this-map)
> 2. [Quick picker](#quick-picker)
> 3. [Pick your core assistant](#pick-your-core-assistant)
> 4. [Research and learning](#research-and-learning)
> 5. [Meetings, writing and presentations](#meetings-writing-and-presentations)
> 6. [Dictation and speed](#dictation-and-speed)
> 7. [Build software and apps](#build-software-and-apps)
> 8. [Automate workflows](#automate-workflows)
> 9. [Images and design](#images-and-design)
> 10. [Video](#video)
> 11. [Voice, audio and music](#voice-audio-and-music)
> 12. [Email, social and sales](#email-social-and-sales)
> 13. [Run AI locally and privately](#run-ai-locally-and-privately)
> 14. [For developers building with AI](#for-developers-building-with-ai)
> 15. [What changed in 2026](#what-changed-in-2026)
> 16. [Before you pay](#before-you-pay)

## How to use this map

Start from the job, not the tool. Answer five questions first:

1. **What's the output?** An answer, a file, a working app, a video, or a workflow that repeats.
2. **Where does the work live?** Google, Microsoft, a code repository, your own computer, or a creative app.
3. **How sensitive is the data?** Public, work-internal, or something that shouldn't leave your machine.
4. **How often will you do it?** Once, weekly, or constantly.
5. **What's the budget?** Free, one subscription, or pay per use.

**Build in layers.** Most people need one core assistant, then a specialist only where it clearly beats the assistant for a job they do often.

| Layer        | What it does                                        | Examples                                  |
| ------------ | --------------------------------------------------- | ----------------------------------------- |
| Core assistant | Thinking, writing, research, files, everyday tasks | ChatGPT, Claude, Gemini, Copilot          |
| Specialist   | One job done better than a general assistant        | Granola, Midjourney, ElevenLabs, Cursor   |
| Glue         | Moves data between tools and runs on a schedule     | Zapier, n8n, scheduled tasks              |

**Read the tables like this:** **Start** shows how you can begin (free plan, free trial, paid, pay per use or open source). **Know this** is the detail worth checking before you commit.

## Quick picker

| I want to...                                  | Start with                          | Also consider                        |
| --------------------------------------------- | ----------------------------------- | ------------------------------------ |
| Ask questions, write and analyze files        | Your core assistant                 | [Pick your core assistant](#pick-your-core-assistant) |
| Get cited answers from the web                | Perplexity                          | Deep research in your assistant      |
| Study my own documents                        | Gemini Notebook                     | Claude or ChatGPT projects           |
| Search academic papers                        | Elicit                              | Consensus                            |
| Take meeting notes                            | Granola                             | Fathom                               |
| Make a presentation fast                      | Gamma                               | Claude Slides, ChatGPT Work          |
| Type by talking, in any app                   | Wispr Flow                          | Superwhisper                         |
| Write and change code with an agent           | Claude Code or Codex                | Cursor, GitHub Copilot               |
| Build an app without coding                   | Lovable                             | Bolt, v0, Replit                     |
| Connect apps and automate a workflow          | Zapier                              | Make, n8n                            |
| Generate images                               | ChatGPT Images or Nano Banana       | Midjourney, Ideogram                 |
| Make logos, icons or vector art               | Recraft                             | Canva                                |
| Generate video clips                          | Google Flow                         | Runway, Kling                        |
| Make a talking-avatar video                   | HeyGen                              | Synthesia                            |
| Edit video or turn long videos into clips     | Descript or CapCut                  | OpusClip                             |
| Create voiceovers                             | ElevenLabs                          | Fish Audio                           |
| Make a song                                   | Suno                                | Udio (no downloads)                  |
| Run AI without sending data to the cloud      | Ollama                              | LM Studio, Jan                       |

## Pick your core assistant

Pick the one closest to where your work already lives. Paying for three rarely helps. Run the same real task in two free tiers and compare.

| Assistant | Pick it when | Paid plans (US) | Know this |
| --------- | ------------ | --------------- | --------- |
| [ChatGPT](https://chatgpt.com) | You want the widest feature set: Work tasks, Codex, Sites, image generation | Go $8/mo, Plus $20/mo, Pro from $100/mo | Free and Go may show ads. See [ChatGPT, End to End](/guides/chatgpt-end-to-end) |
| [Claude](https://claude.ai) | You work with long documents, writing, artifacts, connected apps or code | Pro $20/mo, Max from $100/mo | No image generation. See [Claude, End to End](/guides/claude-end-to-end) |
| [Gemini](https://gemini.google) | You live in Gmail, Docs and Drive | Google AI Plus $4.99/mo, Pro $19.99/mo, Ultra from $99.99/mo | Plans bundle Google storage, Veo video and Nano Banana images |
| [Microsoft Copilot](https://copilot.microsoft.com) | You live in Word, Excel and Outlook | Microsoft 365 Premium $199.99/yr | Includes the Office apps. AI features are for the subscription owner only |
| [Grok](https://grok.com) | You want real-time context from X | SuperGrok tiers | Free to start. One weekly allowance shared across Grok products |
| [Mistral Vibe](https://mistral.ai/products/vibe) | You prefer a European AI provider | Pro $14.99/mo | Formerly Le Chat. Free plan for everyday tasks |

## Research and learning

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Perplexity](https://www.perplexity.ai) | You want fast answers with sources from the web | Free plan | Its Comet browser is free to download |
| [Gemini Notebook](https://notebook.google) | You want answers grounded only in sources you upload | Free plan | Formerly NotebookLM. Existing notebooks carried over |
| [Elicit](https://elicit.com) | You need to search, summarize and compare academic papers | Free plan | Searches more than 138 million papers |
| [Consensus](https://consensus.app) | You want answers drawn from peer-reviewed research | Check pricing | Best for questions research has already studied |
| Deep research in your assistant | You need a multi-source report and already pay for ChatGPT, Claude or Gemini | Included | Open the sources behind any number you plan to reuse |

## Meetings, writing and presentations

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Granola](https://www.granola.ai) | You want AI meeting notes you can search and chat with | Free plan | The free plan limits how much meeting history you can see |
| [Fathom](https://www.fathom.ai) | You want calls recorded, transcribed and summarized | Free plan | Free includes unlimited recordings |
| [Notion AI](https://www.notion.com) | Your team's docs and projects already live in Notion | Free plan | Free includes only a trial of Notion AI. Agents use credits |
| [Gamma](https://gamma.app) | You need a presentation, document or simple site from a prompt | Free credits | Free credits are one-time and don't refresh |
| [Grammarly](https://www.grammarly.com) | You want grammar and tone checks inside the apps you write in | Free plan | Pro has a free trial |

Already on ChatGPT or Claude? ChatGPT Work and Claude Slides and Docs cover many of these jobs without another subscription.

## Dictation and speed

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Wispr Flow](https://wisprflow.ai) | You want to dictate cleanly into any app | Free plan | Also offers a meeting notetaker |
| [Superwhisper](https://superwhisper.com) | You want dictation with offline recognition and custom modes | Free plan | Mac, Windows, iOS and Android |
| [Raycast](https://www.raycast.com) | You want a launcher with AI a keystroke away | Free plan | Raycast AI is part of the paid Pro plan |

Your assistant may already cover this: Claude's quick entry opens with a double-tap of Option on Mac, and ChatGPT's desktop app dictates with `Ctrl+Shift+D`.

## Build software and apps

### Code with an agent

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Claude Code](https://code.claude.com/docs) | You want an agent that works in your repository from the terminal or desktop | Paid Claude plans | Shares usage with the rest of your Claude plan |
| [Codex](https://learn.chatgpt.com/docs) | You want OpenAI's coding agent with reviewable diffs and cloud tasks | Included in ChatGPT plans | Shares usage with ChatGPT Work |
| [Cursor](https://cursor.com) | You want an AI-first code editor | Free plan | Free includes limited agent requests |
| [GitHub Copilot](https://github.com/features/copilot) | Your code lives on GitHub and you want AI in your existing editor | Free plan | Paid plans can run Claude Code and Codex agents |
| [Devin Desktop](https://devin.ai) | You want an editor with long-running autonomous agents | Check pricing | Formerly Windsurf |

### Build apps without code

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Lovable](https://lovable.dev) | You want a full web app from a description | Free plan | Free includes a small daily allowance of build credits |
| [Bolt](https://bolt.new) | You want to prototype and host a site or app quickly | Free plan | Free sites carry Bolt branding |
| [v0](https://v0.app) | You want web interfaces that deploy to Vercel | Free plan | Free includes a daily message limit |
| [Replit](https://replit.com) | You want to build, run and host in one browser workspace | Free plan | Paid plans add more agent usage and collaborators |

Before you build a whole app, check whether a [ChatGPT Site](/guides/chatgpt-end-to-end#images-visuals-and-sites) or a [Claude artifact](/guides/claude-end-to-end#artifacts-visuals-and-designs) already does the job.

## Automate workflows

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Zapier](https://zapier.com) | You want the most app integrations with the least setup | Free plan | Free includes 100 tasks a month and two-step workflows |
| [Make](https://www.make.com) | You want visual, branching workflows with more control | Check pricing | Compare what counts as a billable step before you build |
| [n8n](https://n8n.io) | You want workflows you can self-host and extend with code | Free trial | Can be self-hosted. Cloud plans charge per workflow run, not per step |
| [Manus](https://manus.im) | You want an agent that plans and carries out a whole task | Free trial | Meta's acquisition was blocked and is being unwound. The product is still live |
| [Genspark](https://www.genspark.ai) | You want one agent workspace for research, slides, docs and calls | Check pricing | Compare credit costs before running long agent tasks |
| Scheduled tasks in your assistant | The job is a recurring brief or check you've already prompted well | Included | Covered in both [ChatGPT](/guides/chatgpt-end-to-end#scheduled-tasks-and-long-running-work) and [Claude](/guides/claude-end-to-end#scheduled-tasks-and-long-running-work) guides |

Not sure a schedule is the right shape? Read [Routine vs Schedule vs Loop](/guides/routine-vs-schedule-vs-loop).

## Images and design

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| ChatGPT Images | You want to generate and edit images in conversation | ChatGPT plans | Image turns use your limits faster than text |
| [Nano Banana](https://gemini.google) | You're in Gemini or Google Flow and want fast image edits | Limited free use | Paid Google AI plans include more credits |
| [Midjourney](https://www.midjourney.com) | Visual style and aesthetic quality matter most | Check pricing | Can also animate images into short video clips |
| [Ideogram](https://ideogram.ai) | You need readable text inside the image | Check pricing | Proofread every word anyway |
| [Recraft](https://www.recraft.ai) | You need vectors, icons or consistent brand illustrations | Free plan | Images made on the free plan are public and owned by Recraft |
| [Krea](https://www.krea.ai) | You want many image and video models in one place | Free plan | Free includes a small daily allowance |
| [Adobe Firefly](https://firefly.adobe.com) | You already work in Adobe apps | Check plans | Includes 30+ models across image, video and audio |
| [Canva](https://www.canva.com) | You need finished social posts, docs or decks from templates | Free plan | Brand kits and more AI features are on paid plans |
| [Figma](https://www.figma.com) | You design product interfaces with a team | Free plan | Paid seats include monthly AI credits |

On a Claude paid plan, **Claude Design** builds mockups and one-pagers that follow your design system.

## Video

### Generate footage

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Google Flow](https://flow.google.com) | You want cinematic clips with Google's Veo models | Google AI Plus and up | Uses credits from your Google AI plan |
| [Runway](https://runwayml.com) | You want several video models plus editing tools in one app | Free credits | Free is a one-time 125 credits |
| [Kling](https://kling.ai) | You want text- or image-to-video with consistent multi-scene shots | Check pricing | Also available inside Runway |
| [Higgsfield](https://higgsfield.ai) | You want video, animation and ad generation across many models | Check pricing | Compare credit costs per model before a big project |

### Avatars and presenters

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [HeyGen](https://www.heygen.com) | You want a talking avatar or a translated version of your video | Free plan | Free includes 3 videos a month, up to 1 minute |
| [Synthesia](https://www.synthesia.io) | You make training or internal videos at scale | Free plan | Paid plans add custom avatars and downloads |

### Edit and repurpose

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Descript](https://www.descript.com) | You want to edit video or podcasts by editing the transcript | Free plan | Paid plans remove watermarks and add AI credits |
| [CapCut](https://www.capcut.com) | You make short-form social videos with captions and templates | Free plan | Works on desktop, mobile and the web |
| [OpusClip](https://www.opus.pro) | You want to turn long videos into short clips | Free plan | Free includes 60 credits a month |

## Voice, audio and music

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [ElevenLabs](https://elevenlabs.io) | You need voiceovers, voice cloning, dubbing or sound effects | Free plan | Commercial use starts on paid plans |
| [Fish Audio](https://fish.audio) | You want expressive text-to-speech and voice cloning | Free plan | Check the license for each voice you use |
| [Suno](https://suno.com) | You want full songs from a prompt | Free plan | Commercial use rights come with paid plans |
| [Udio](https://www.udio.com) | You want to make and share music inside Udio | Check pricing | Downloads are disabled, so tracks can't be used elsewhere |
| [Adobe Podcast Enhance](https://podcast.adobe.com/enhance) | You need to clean up voice recorded in a noisy room | Free | Works on speech, not music |

Only clone a voice you own or have permission to use.

## Email, social and sales

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Kit](https://kit.com) | You run a creator newsletter | Free plan | Free covers up to 10,000 subscribers |
| [Manychat](https://manychat.com) | You want to automate Instagram, Facebook or WhatsApp DMs | Check pricing | Follow each platform's messaging rules |
| [Clay](https://www.clay.com) | You research and enrich lead lists from many data sources | Free plan | Free includes 500 actions a month |

## Run AI locally and privately

Local models keep data on your machine and cost nothing per message. They're usually less capable than frontier models and need a reasonably powerful computer.

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [Ollama](https://ollama.com) | You want the simplest way to run open models from the command line or apps | Free locally | A paid plan adds cloud usage |
| [LM Studio](https://lmstudio.ai) | You want a desktop app to download and chat with local models | Check pricing | Now includes Bionic, an agent for documents, slides and code |
| [Jan](https://www.jan.ai) | You want an open-source ChatGPT-style app for local or cloud models | Free, open source | Connects to cloud models too, so check which one you're using |
| [Open WebUI](https://openwebui.com) | You want a self-hosted chat interface for a team or home server | Open source | You run and secure the server yourself |

Check each model's license before using it commercially.

## For developers building with AI

| Tool | Pick it when | Start | Know this |
| ---- | ------------ | ----- | --------- |
| [OpenRouter](https://openrouter.ai) | You want one API key for models from many providers | Pay per use | Buy credits once and use them with any model |
| [Hugging Face](https://huggingface.co) | You need open models, datasets or a place to host demos | Free plan | Pro is $9/month |
| [Replicate](https://replicate.com) | You want to run open image, video or language models by API | Pay per use | Some models bill by compute time, others by input and output |
| [fal](https://fal.ai) | You need fast image and video generation APIs | Pay per use | Also rents GPUs by the hour |
| [AI SDK](https://ai-sdk.dev) | You're building AI features in TypeScript | Open source | Works with 100+ models |
| [Langfuse](https://langfuse.com) | You need tracing, evaluations and prompt management | Free plan | Can be self-hosted |

## What changed in 2026

If a tutorial names one of these, it may be out of date.

| Tool | What changed | When |
| ---- | ------------ | ---- |
| Sora | OpenAI shut down the app and website. The API ends September 24, 2026 | April 26, 2026 |
| Windsurf | Renamed Devin Desktop after Cognition's acquisition | June 2, 2026 |
| NotebookLM | Renamed Gemini Notebook. Same product, notebooks carried over | July 16, 2026 |
| Le Chat | Now Mistral Vibe | 2026 |
| Claude Cowork | Merging with chat into one Claude conversation, starting with Pro and Max | Rolling out |
| GPT-5.5 | Retiring from ChatGPT, ChatGPT Work and Codex | October 14, 2026 |
| Manus | China blocked Meta's acquisition. The deal is being unwound and the product is still live | April 27, 2026 |
| Udio | Downloads disabled after its Universal Music settlement | October 30, 2025 |

## Before you pay

Run through this list before you add another subscription:

1. **Does your core assistant already do it?** Try the job there first.
2. **Can you get your work out?** Check exports, downloads and file formats.
3. **Who owns the output?** Some free plans make your images public or exclude commercial use.
4. **Will your data be used for training?** Find the setting before you upload anything sensitive.
5. **Subscription or credits?** Credits can run out fast on video, agents and retries.
6. **Can your team use it?** Check admin controls, shared billing and security reviews.
7. **Test it on a real task.** Run the same job in two tools and keep the one whose output you'd actually use.

**Go deeper:** [ChatGPT, End to End](/guides/chatgpt-end-to-end) · [Claude, End to End](/guides/claude-end-to-end) · [ChatGPT Work vs Codex](/guides/chatgpt-work-vs-codex) · [How to Give AI Taste](/guides/how-to-give-ai-taste)