Get started
Page 1 of 6
Quickstart
Install the app, send your first request, and see where it ran. About five minutes.
Charon gives every app on your computer one AI endpoint. Each request runs on your own machine when it can, or in a cloud you set up when it should.
1. Install the app
Download Charon for free. It needs a Mac with Apple silicon and macOS 13 or later (Linux, Windows and Intel Macs are coming soon). Open it and follow the setup: it starts Charon and offers a model that fits your Mac.
2. Point your app at Charon
Use the base URL http://127.0.0.1:7777/v1 and the model auto. Apps on your Mac need no key, so any
placeholder works.
from openai import OpenAI
client = OpenAI(base_url="http://127.0.0.1:7777/v1", api_key="charon")
reply = client.chat.completions.create(
model="auto",
messages=[{"role": "user", "content": "Hello!"}],
)
print(reply.choices[0].message.content)curl http://127.0.0.1:7777/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{"model": "auto", "messages": [{"role": "user", "content": "Hello!"}]}'Streaming, tool calls, JSON mode, embeddings and the Responses API work as they do with OpenAI.
3. See where it ran
The app’s dashboard lists every request, where it ran, and why. Each response also says so in its headers:
X-Charon-Backend: local
X-Charon-Model: qwen2.5-7b-instruct
X-Charon-Locality: local
X-Charon-Reason: local: lowest latency (est 180 ms), cost 0
X-Charon-Request-Id: 9f2c…Optional: the command line
The app includes the charon command. To use it in a terminal:
sudo ln -sf /Applications/Charon.app/Contents/Resources/charon /usr/local/bin/charon
charon statusNext: add a cloud in Backends, or set budgets and local-only rules in Routing and policies.