AI that works where your data lives.
Run AI on your own machines, or connect your AI tools through the CLI and MCP. By default, your message content stays on your machine.
1
What breaks
Many teams cannot send patient data to cloud AI, so the work has stayed manual.
2
Why the usual fix falls short
A dev box with synthetic data can use any model. A machine next to live channels must keep message content on your network.
3
What Pidgeon does
On a dev box, your AI tools drive Pidgeon through the CLI and MCP. Inside the apps, the built-in assistant runs on a local model by default, or on an endpoint you choose.