Does livekit agents support minimax m2.7 as llm mo...
# ask-aia
ancient-dog-5313
03/30/2026, 9:50 PMDoes livekit agents support minimax m2.7 as llm model?
✅ 1
t
tall-airline-16756
03/30/2026, 9:50 PMThanks for reaching out, edmon. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
tall-airline-16756
03/30/2026, 9:51 PMThanks — you're asking whether LiveKit Agents supports the Minimax M2.7 LLM.
Short answer: No. LiveKit Agents does not support Minimax M2.7 as an LLM. Minimax is available only as a Text-to-Speech (TTS) provider in LiveKit (via the MiniMax TTS plugin), and Minimax is not listed among the supported LLM providers or accessible through LiveKit Inference or an LLM plugin for agent use.
If your goal is to use Minimax M2.7 specifically as the model behind agent responses, a couple of options you might consider:
• Use a different LLM provider that LiveKit supports (OpenAI, Google Gemini, Anthropic, Mistral, Groq, etc.) for agent text generation.
• If you control an API endpoint for Minimax M2.7, tell me more about your setup — I can help explore whether creating a custom LLM plugin or proxy is feasible with your API so the agent can call that model.
Which path would you like help with: switching to a supported LLM, or exploring a custom integration for Minimax?
Sources: OpenAI LLM models | LiveKit Documentation | Large language models (LLM) overview | LiveKit Documentation | MiniMax TTS plugin guide | LiveKit Documentation | Models overview | LiveKit Documentation | Baseten LLM plugin guide | LiveKit Documentation, with interactive elements
a
ancient-dog-5313
03/30/2026, 9:51 PMcan I use it as a openai compatible model?
t
tall-airline-16756
03/30/2026, 9:52 PMThanks — you're asking whether you can present Minimax M2.7 to LiveKit Agents as an OpenAI-compatible model.
Short answer: You cannot natively treat Minimax M2.7 as an OpenAI model in LiveKit Agents. LiveKit Agents expects OpenAI-compatible endpoints only when using an OpenAI-style provider; Minimax is not listed as an OpenAI-compatible provider and LiveKit does not include built-in support to map Minimax M2.7 directly to the OpenAI provider.
That said, you have two practical paths to achieve similar behavior:
• Use a custom LLM plugin or proxy that translates OpenAI-style calls to Minimax’s API.
◦ If Minimax exposes an HTTP API, you can run a small adapter service that accepts OpenAI-like requests and forwards them to Minimax (mapping parameters, streaming, and responses as needed).
◦ Then configure LiveKit Agents to use that adapter as if it were an OpenAI-compatible endpoint (or register it as a custom plugin if your setup supports that).
◦ I can help outline the adapter behavior and required request/response mapping if you can share Minimax’s API format (endpoints, auth, streaming vs non-streaming).
• Use a supported LLM provider instead.
◦ If you want a simpler integration with no middleware, choose one of the LLM providers LiveKit already supports (OpenAI, Google Gemini, Anthropic, Mistral, Groq, etc.).
Which would you like to pursue?
• If you want the adapter/proxy approach: please share Minimax’s API details (authentication method, request body format, and whether it supports streaming). I’ll help draft the adapter mapping and required config for LiveKit Agents.
• If you prefer switching providers: tell me which supported provider you’re considering and I’ll outline the steps to configure it with LiveKit Agents., with interactive elements
2 Views