Join waitlistRequest early access
← Back to home
Product · Developers

Your local models, behind one API.

Nupick exposes your local models behind an OpenAI-compatible API. Point any OpenAI client at your own machine — no data leaves unless you choose to route to the cloud.

bash
# Point any OpenAI client at your local Nupick
curl http://localhost:48420/v1/chat/completions \
  -H "Authorization: Bearer $NUPICK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen2.5:7b",
    "messages": [{"role": "user", "content": "Hello from my own machine."}]
  }'
OpenAI-compatible
GET /v1/models and POST /v1/chat/completions speak the OpenAI shape. Most OpenAI SDKs work by just changing the base URL.
Key-gated, off by default
The API is disabled until you turn it on and generate a key — when it's off, the surface simply isn't there.
Serve on your LAN
Opt in to expose the endpoint to other devices on your network — your own private inference server, no cloud account.
Extend with MCP
Add your own Model Context Protocol tools and resources. Nupick is built on the open standard, not a closed plugin list.
Local by default

A private inference server that happens to have a UI.

Requests served by the local API run on your machine. Nothing crosses to the cloud unless you explicitly route a task there — and when it does, the same privacy boundary applies.

Early access

Start building your
private AI workspace.

Local-first, privacy-first, and built for people who want powerful AI without handing over everything.

Request early access →Contact the team