Chat & models

Send chat messages and discover which models your workspace can use.

Note

All requests require a Bearer API key. See Authentication. Base URL: https://ai.mixerlead.com/api.

List models

GET /api/ai/models

Returns the models available to your workspace’s plan. Use the returned model id values with the chat endpoint.

curl https://ai.mixerlead.com/api/ai/models \
  -H "Authorization: Bearer $MIXERLEAD_API_KEY"

Response (200)

{
  "models": [
    { "id": "ml-auto", "name": "Auto", "task": "Automatic model selection", "requiresPaid": false },
    { "id": "<model-name>", "name": "<model-name>", "task": "Text generation", "requiresPaid": false }
  ]
}

Tip

Use ml-auto (the default) to let MixerLead pick a strong model for the task. Specify a model only when you need consistent behavior. Models marked requiresPaid need a paid plan.

Send a chat message

POST /api/chat

Sends a message and returns the assistant’s reply. The response is streamed so you can render tokens as they arrive.

Request body

Field	Type	Required	Description
message	string	✅	The user message to send.
model	string	—	A model id from /api/ai/models. Omit for Auto.
conversationId	string	—	Continue an existing conversation. Omit to start a new one.

curl -N https://ai.mixerlead.com/api/chat \
  -H "Authorization: Bearer $MIXERLEAD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Write three LinkedIn angles for our product launch.",
    "model": "ml-auto"
  }'

Node.js — read the stream

const res = await fetch("https://ai.mixerlead.com/api/chat", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.MIXERLEAD_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ message: "Summarize our Q3 strategy in 5 bullets." }),
});

// Stream the response body as it arrives.
const reader = res.body.getReader();
const decoder = new TextDecoder();
for (;;) {
  const { value, done } = await reader.read();
  if (done) break;
  process.stdout.write(decoder.decode(value));
}

Streaming

/api/chat returns a streamed response. Read the body incrementally (as above) rather than waiting for a single JSON object. The -N flag disables curl’s buffering so you see tokens live.

Grounding

Chat sent through the API is grounded in your brand profile automatically. To ground answers on your own documents, add sources via the Knowledge API first.

Limits & credits

Chat consumes credits and counts against your plan’s rate limits. Handle 429 responses with backoff — see Errors & rate limits.

Related

Skills & agents Run templated skills and agents.

Knowledge Add sources for grounded answers.

Errors & rate limits Status codes and retries.