Learn AI Free
Free AI courses, tutorials, and guides. Learn to use AI effectively.
Start here: your first 5 minutes
- Pick a tool - Chat, Image, Voice, Video, or any of the 400+ tools. No account needed to try.
- Type what you want in plain language. There is no special syntax - "make me a logo for a coffee shop called Bloom, minimal, green" is a perfect prompt.
- Iterate. Your second prompt is always better than your first - say what to change ("same but warmer light", "shorter, friendlier tone") instead of starting over.
- Create a free account when you hit the visitor limit - confirmed accounts get 30,000 free tokens every day, 12x the anonymous allowance. Everything you make is yours, commercial use included.
Prompting fundamentals (works on every model)
1. Say what, for whom, and in what format
Weak prompts are vague about the output. Strong prompts name the audience, the format, and the constraint.
Weak: “write about our product”
Strong: “write a 3-paragraph product announcement for existing customers, friendly but not salesy, ending with one clear call to action”
2. Give context the model can't guess
Paste the email you're replying to, the code that's failing, the paragraph you want rewritten. Models are much better at transforming what you give them than at inventing what you meant. In Chat you can also attach files - images, audio, PDFs - and ask questions about them.
3. For images: subject, style, lighting, composition
Image models respond to concrete visual language. A reliable recipe: subject + setting + style + lighting + framing.
“a red fox in a snowy forest, watercolor style, soft morning light, close-up portrait” - then use the negative prompt for what you don't want (“blurry, text, watermark”)
4. Compare models when quality matters
Different models have different strengths. Run the same prompt on up to 4 models side-by-side in chat compare or image compare and keep the winner.
Tool playbook
- Chat - it can also DO things: say "make me an image of…", "search the web for…", or "say this out loud" and it calls the right tool mid-conversation.
- Image - start with the free SDXL model (~1,000 tokens). Need a specific pose or layout? ControlNet follows a reference image.
- Voice - 174 voices in 37 languages. The live cost box quotes your exact price as you type. Long scripts: turn on long-form stitching.
- Video - clips cost 15,000 tokens minimum, then 2,500 per second beyond 6 seconds. Free accounts get 1 clip/day; for a multi-scene story use the Movie tool.
- Transcription - upload audio or video, get timestamped text with speaker labels. About 500 tokens per minute of audio.
- Translate - hundreds of languages, plain text or whole documents.
- Music - describe genre, mood, and instruments (“lo-fi beat, warm, vinyl crackle”). A song is ~5,000 tokens; you can split any track into stems after.
- Coder - an in-browser IDE with an AI agent that edits files and runs commands. Ship it to a real VPS from the same screen.
How tokens work
One balance covers every tool - you only spend when you generate. Confirmed free accounts get 30,000 tokens every day, which covers everyday use; heavy creative tools have small daily caps on free accounts (5 images, 2 songs, 1 video) that any token pack or plan removes.
Chat reply ≈ 500-2,000 Image ≈ 1,000 tokens Song ≈ 5,000 Video ≈ 15,000 min/clipFor developers
Everything on the site is one OpenAI-compatible REST API. Generate a key and your first call works in under 2 minutes - token costs match the web exactly.
API Docs Get API KeyModel Directory
Every model - self-hosted open-source and premium - with per-use token costs, strengths, and a working generator on each page.
Browse Models Model ArenaMore Resources
- Open Source Models - every self-hosted model with its license and source link
- Integrations - Slack, Zapier, WordPress, the embeddable widget, and more
- AI Workflows - chain tools together for multi-step automation
- Community Gallery - see what others made, remix any prompt with one click
FAQ
A 5-minute quick start, prompting fundamentals with concrete before/after examples, a per-tool playbook (chat, image, voice, video, transcription, music, code), an honest token cost guide, and the API quickstart for developers.
Pick any of the 400+ tools and type what you want - no signup needed to try. Creating a free account and confirming your email unlocks 30,000 free tokens every day (12x the visitor allowance). For developers, generate an API key to integrate AI into your own apps.
Yes - the tool playbook above covers the main categories, and every tool page has its own how-to section and 12-question FAQ tailored to that tool.
Full API documentation is at free.ai/api/ - endpoints, authentication, request/response formats, and code examples in Python, JavaScript, and cURL. The API is OpenAI-compatible, so most existing SDKs work by changing the base URL.
No. The web interface is designed for non-technical users - type your request in plain language and the AI responds. Technical users can use the API for programmatic access, but it is not required for any of the 400+ tools.
Start with the free self-hosted default in each tool - it's good for most tasks and costs the least. If the result isn't good enough, run the same prompt through compare mode or step up to a premium model; the model picker shows each option's token cost so you can judge the trade.
Yes - see the prompting fundamentals section above. The short version: say what, for whom, and in what format; paste the material you want transformed; use concrete visual language for images; and iterate by describing the change instead of starting over.
Yes. The open-source page lists every self-hosted model with its license, use case, and source link. All our self-hosted models use permissive licenses - Apache 2.0, MIT, or BSD - and everything you generate is yours, commercial use included.
The integrations page links to setup guides for Slack, WordPress, Zapier, the embeddable widget, and the bookmarklet. Each guide includes step-by-step instructions and code examples.
Tokens are the universal currency on Free.ai - one balance covers every tool, and you only spend when you generate. Confirmed accounts get 30,000 free tokens daily. Typical costs: a chat reply 500-2,000, an image ~1,000, a song ~5,000, video 2,500 per second. See the tokens page for the full breakdown.
The community gallery is the best place to see and remix what others are making. For help, use the contact form or the detailed documentation on each tool page.