I'm Daniel, part of the team at Agnes AI, a Singapore-based AI lab. We've been building quietly for a while and I wanted to share what we've made with this community and get honest feedback from people who actually build things.
What is Agnes AI? Agnes AI gives you access to three models through a single free API. We have text, image and video models. The API is fully OpenAI-compatible, so you can just swap the base URL to https://apihub.agnes-ai.com/v1 and use your Agnes API key. No other changes needed in your existing code. It is completely free. No need to submit any kind of payment information. After registration there will be a default balance so that you can create your API keys. You sign up, get a key, and start building. Thanks to the community, we hit 4T API calls in just one week, and it keeps growing. We are also ranked as a top 10 AI lab on benchmarks. Our biggest advantage is that the free API is fully multimodal, so you can use text, image and video generation all without limits.
The models: Text - Agnes-2.0-Flash: It has 512k context window (in a week or two it will be 1M). There are reasoning/thinking modes, tool calling, image URL input and streaming support. Ranked 9th in general tasks on Claw-Eval. Can be used with any coding assistant. I personally use it with OpenCode and Hermes. Can do automation and agent tasks, large context tasks, and reasoning tasks. And of course can be used for just general tasks. For the automation the community is already integrating with n8n and doing some automations from telegram for free. Landing page generated with Agnes: https://codebuddy.work/agents/share/H4zVVxaFpU2I2UDcVSJAg-MH...
Image - Agnes-Image-2.1-Flash: Text-to-image and image editing available. It can generate up to 4k images. For the rankings also can check Artificial Analysis image editing leaderboard. Here is a 4K image generated: https://imgur.com/LkMHnr7 and https://imgur.com/a/0UsRiGX
Video - Agnes-Video-V2.0: Text-to-Video and Image-to-Video. Also supports multiple images to video. Ranked on Artificial Analysis video leaderboard.
Why free? We want developers to actually use it and tell us what is broken, what is missing, and what works well. That feedback matters more to us right now than revenue. Also our mission is to make AI accessible to all.
Things people have already built with it: A custom node pack for ComfyUI connecting Agnes image and text models, a coding agent skill supporting all Agnes 2.0 models (https://github.com/lj1270998580-crypto/Agnes-help-skill), Hermes integration for agentic local workflows (https://x.com/decapostos/status/2065758852738416779) and much more.
Links: The official website: https://agnes-ai.com/
The API platform: https://platform.agnes-ai.com/
GitHub: https://github.com/AgnesAI-Labs/Agnes-AI (can check the model updates)
Docs: https://agnes-ai.com/doc
Discord: https://discord.gg/JAWe3YVYjr
Text model benchmarks: https://claw-eval.github.io
Image model leaderboard: https://artificialanalysis.ai/image/leaderboard/editing
Video model leaderboard: https://artificialanalysis.ai/video/leaderboard/image-to-vid...
Happy to answer any questions about the models, the API, or how we built things. Would love to hear what you build with it. Feel free to join our Discord community too, lots of people sharing what they've made there.
Thanks