Local server (Hrida AI Studio)
The desktop app can install and run the full Hrida AI Studio server on your machine, then connect to it as a local connection. Managed entirely from Settings → Hrida.ai — no terminal needed.
One-click install
When Studio isn't installed yet, Settings → Hrida.ai shows an Install button. Behind it, the app:
- Downloads a standalone Python runtime (python-build-standalone) into the install directory.
- Installs the
uvpackage manager. - Installs the bundled
hrida-ai-studiowheel shipped inside the app (it's a private/commercial package, not on any public index).
The Python runtime and packages live under the configurable install location (Settings → General → Advanced); server data lives under the data location (below).
Running the server
| Control | Effect |
|---|---|
| Start / Stop / Restart | Manage the server process. Status shows running / stopped and the version. |
| Start on launch | Automatically start the server when the app opens. |
| Server port | Which port the server binds (default 8080). |
| Serve on local network | Bind beyond localhost so other devices on your network can reach this server. Off by default. |
| Auto-update | Keep the bundled server package up to date automatically. |
| Update | Manually update the server package now. |
The running instance panel shows the live URL and process info, and the server's logs are viewable in-app (with a clear-logs action). A PTY view lets you attach to the server process output directly.
Data location
Server data (database, uploads, config) defaults to a per-user application-data path. Override it in Settings → Hrida.ai → Data location — browse to a folder or type a path. Changing it points a fresh server at the new location.
Uninstall
Settings → Hrida.ai → Uninstall removes the hrida-ai-studio package. Removing the Python runtime itself is handled separately (see Settings).
Environment variables
Custom environment variables set in Settings → General → Advanced → Environment variables are passed to the server (and other managed child processes) when they start — useful for provider API keys, proxies, or feature flags.