Features that help you get more done
From a quick chat to a multi-step flow: see what Meteor-AI does for you. Every feature is there from day one, even on the free plan. App status: September 2026.
Chat with the model of your choice
About 390 models in one chat – the right one for every question. Switch whenever you like and see up front what each one costs.
- Streaming: answers appear word by word and can be stopped at any time.
- Model picker: search by name, context size, capabilities and price per million tokens (Ctrl+K).
- Markdown: code blocks with a copy button, tables, lists and links.
- History: search, rename and keep chats in folders.
- Controls: keyboard shortcuts, light or dark, adjustable chat width – usable from 320 pixels.
Web search and thinking
Current facts with sources and well-reasoned answers to hard questions – with two switches in the plus menu. Once on, they stay on until you switch them off.
- Web search: the model reads current sources and cites them under the answer – compact or thorough.
- Thinking: the model thinks before it answers; the reasoning sits in a collapsible block above. Six effort levels.
- Always a fit: if a model can’t do something, the switch is disabled and says why.
Attach files and images
Contracts, spreadsheets, screenshots: just attach them and the model works with them directly. No more retyping or copying.
- PDF: read in full, with no length limit. For very long documents a notice points out the cost.
- Text, code, CSV, JSON: placed as a code block before your message.
- Images: up to 4 MB, for every model that can see. Scanned PDFs can be attached this way too.
Create and edit images
Create images as easily as having a conversation: say what you want and refine it round by round – the last image is always the template for the next.
- From text: describe the image, pick an image model, generate.
- Step by step: “make the sky clearer” changes the existing image instead of inventing a new one.
- Your own template: an image from your computer (up to 5 MB) is the starting point.
- Style from the library: a skill such as “35 mm, soft light” doesn’t need typing every time.
- Saved: all images are kept; a grid view shows them at a glance.
Assistants: your AI specialists
An assistant is a ready-configured chat for one task – with model, role, skills and knowledge. Set it up once and it delivers answers by your rules every time.
- Configurable: model, system prompt, creativity, thinking and web search, plus name, icon and tags.
- Skills and knowledge: assigned from the library and sent with every message.
- Conversation starters: up to three opening questions for a quick start.
- Own conversations: chats with an assistant stay with it.
Flows: several AI steps, one input
Let several models work together for you: a flow links AI steps, each with its own model and prompt. One input is enough and the whole routine runs.
- In sequence: each step builds on the previous one – you only see the result.
- In parallel: all steps work on the same input side by side, each delivering its part.
- Per step: model, prompt, creativity, thinking, web search – and where the output goes.
- As a diagram: the editor shows the process as a picture.
Library: build once, use forever
Your best know-how in one place: seven areas for everything you need again and again – each with tiles, search and tag filter.
- Skills and prompts: templates for tasks, questions and the model’s role.
- Knowledge items: texts and documents as context for answers; PDFs are read out.
- Create with AI: have system prompts and skills written from a description and refine them in the same window.
- In the chat: attach a prompt, skill or knowledge to a single message via the plus menu.
HTML apps: tools without coding
A calculator, a checklist, a form: describe in chat what you need and AI builds the app. Store it in the library and launch it whenever you need it.
- Ready to run: preview in the editor, launch in its own window or straight from the tile.
- Isolated: the app runs sandboxed, with no access to your account, cookies or storage.
- No server: everything lives in one document – nothing to set up.
Costs, plan and statistics
No surprises on your bill: every request is checked beforehand, every answer booked. You always see what you’ve used – and never pay more than is covered.
- Price up front: every model shows its price per million tokens.
- Cost per answer: tokens and cost appear in the header of every answer.
- Always covered: before every request Meteor-AI checks allowance and credit; if it runs low, the answer gets shorter.
- Plan & payment: plan, allowance and credit at a glance, payment via Stripe.
- Statistics: usage by period with CSV export – without prompts or answers.
Account, data and help
An account you control yourself – from sign-up to deletion.
- Sign-in: registration with a confirmation email, password reset by link.
- Storage at a glance: an overview shows how much space attachments, images and knowledge take up.
- Delete: all data or the whole account; remaining credit is refunded.
- Help in the app: guides, a model table with prices and a support form.
See for yourself – for free.
Start for free and test every feature, no payment method needed.