SOL 1 SERIES / NOW AVAILABLEREAD THE RELEASE NOTES
mindlarp_
HTTP API

A small, familiar surface.

MindLarp uses an OpenAI-compatible chat completions interface over JSON and HTTPS.

Base URL

For this local application, the base URL is http://localhost:3000/v1. Use your public HTTPS origin in production.

List models

GET/v1/models

Requires the models:read scope. Returns the available model catalog.

Create chat completion

POST/v1/chat/completions

Requires the inference:write scope.

Body parameters

FieldTypeDescription
modelstring · requiredA MindLarp model identifier.
messagesarray · requiredMessages with role and content.
streambooleanStream the upstream response when supported.
temperaturenumberSampling temperature, from 0 to 2.
max_tokensintegerMaximum output tokens, from 1 to 32768.
Request
{
  "model": "MindLarp-Sol-1-Plus",
  "messages": [
    {"role": "system", "content": "Be concise."},
    {"role": "user", "content": "Why is the sky blue?"}
  ],
  "temperature": 0.7
}

Inference configuration

Set MINDLARP_UPSTREAM_URL to your OpenAI-compatible model service and MINDLARP_UPSTREAM_KEY to its credential. Without an upstream, the endpoint returns a clear 503 inference_not_configured error instead of a fabricated answer.

Errors

Errors use a consistent JSON object with a message, type, and optional machine-readable code. See Errors & limits.