Set up the Local API Server

The Local API Server lets your own scripts and tools talk to Tholos AI through an OpenAI-compatible API — running on your machine, localhost only. It’s how you automate or integrate the on-device engine without anything ever leaving your device. This is a Business edition feature.

Before you start

• You’re on the Business edition (the setting is locked otherwise).
• A model is loaded in the app — the server only starts serving once there’s a model to answer with.

Turn it on

1. Open Settings → Local API Server.
2. Switch on Enable Local API Server.
3. The Status line shows where it’s listening, e.g. Running — http://127.0.0.1:8756. (If no model is loaded yet, it reads “Enabled — starts once a model is loaded.”)

Get your API key

Every request must be authenticated with the key shown in the API Key row:

• Copy — copies the key to your clipboard. Treat it like a password.
• Regenerate — issues a new key and invalidates the old one (any script using the old key stops working until you update it).
• Test — runs a quick built-in check and reports whether the server answered.

Your connection details

• Base URL: http://127.0.0.1:8756/v1 (use the port shown in your Status line)
• Auth header: Authorization: Bearer YOUR_API_KEY

Make your first call

With the server running, this sends a chat request from a terminal:

curl http://127.0.0.1:8756/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

Omitting model uses whichever model is currently loaded in the app. For the full endpoint list, parameters, and code samples in Python and JavaScript, see the Local API Server reference.