Can I add Mistral as a custom LLM for my voice agents?

[lament flower]

AI-generated Answer for yves_69297

Yes. Mistral isn’t a native provider in Agents, but you can use it via Custom LLM as long as you point your agent to an OpenAI chat-completions–compatible endpoint. (elevenlabs.io)

How to set it up

• Add your LLM API key as a Secret, then select Custom LLM and enter the Server URL (your OpenAI-compatible endpoint), Model ID (e.g., a Mistral/Mixtral model), and the secret; enable “Custom LLM extra body” if prompted. (elevenlabs.io)
• Use a model that supports tool use/function calling so Agents features (system tools) work correctly. Together AI’s guide lists Mistral/Mixtral models that support function calling. (elevenlabs.io)

Notes

• When you choose a Custom LLM, ElevenLabs’ LLM cascading/fallback is disabled; the system will retry only your custom endpoint. (elevenlabs.io)
• For general capabilities: Agents let you pick a hosted model or bring your own custom LLM. (help.elevenlabs.io)

Sources:
Models | ElevenLabs Documentation
Integrate your own model | ElevenLabs Documentation
Together AI | ElevenLabs Documentation
LLM Cascading | ElevenLabs Documentation
What can I create with ElevenLabs Agents (formerly Conversational AI) – ElevenLabs

[devout kayak]

Do you know what requirements I need to self-host a LLM capable enough to call tolls?