Execute 150+ AI models for images, video, audio, LLMs, and 3D generation via a single HTTP endpoint. Ideal for developers building multi-modal AI workflows without managing separate model infrastructure.
ac.inference.sh/mcp
https://sh.inference.ac
HTTP
No auth required
How models use it and what it is built for.
Execute 150+ AI models for images, video, audio, LLMs, and 3D generation via a single HTTP endpoint. Ideal for developers building multi-modal AI workflows without managing separate model infrastructure.
Hosted endpoint — paste into any MCP client.
Where to find authoritative docs and source for inference.sh.
Paste any of these into Agent Studio after connecting inference.sh.
Common questions about connecting and running inference.sh.
What AI models does this MCP server support?
The server provides access to 150+ AI applications spanning image generation, video synthesis, audio processing, LLMs, and 3D model creation. You can browse available models through the MCP interface to see the full catalog.
How do I authenticate and set up the server?
The server is hosted at https://sh.inference.ac and uses HTTP transport. Check the environment variables and authentication requirements in your MCP client configuration; specific auth details should be in your setup documentation.
Can I stream results from long-running inference tasks?
Yes, the server supports streaming results, which is useful for video, audio, and large model outputs. This allows you to receive partial results as they complete rather than waiting for the full response.
What's the difference between this and running models locally?
This MCP server provides hosted inference, so you don't need to manage GPU infrastructure, download model weights, or handle deployment—just call the endpoint. Trade-off: you depend on external service availability and may incur usage costs.
How do I know which model to use for my task?
Use the browse capability in the MCP server to list available models and their descriptions. Each model typically includes input/output schema details so you can match it to your use case (e.g., image generation vs. video synthesis).
Connect inference.sh to Claude, GPT, Gemini, DeepSeek and 30+ AI models in MCP Agent Studio. Compare answers side-by-side, save reusable agent presets, share runs — all in your browser, no install required.
Open Agent Studio