Run dashboards locally
Download a dashboard your AI employee built and run it on your own machine with the AI Employee Runner — no Docker, no tunnel, no VPN into the cloud.
Why run locally
Some dashboards need data that only your machine can reach: a database on your company VPN, a local file, or a service on localhost. A cloud container cannot see those without building a tunnel or a VPN bridge.
Running the dashboard on your own machine flips this around. The dashboard runs where the data already lives, so it connects directly. Nothing is exposed to the public internet and no inbound tunnel is needed.
Run a dashboard locally
- 1In a dashboard preview, click "Run Locally" to download a .aiemp bundle (a small zip with the dashboard and a manifest — no data, no secrets).
- 2Install the AI Employee Runner once: a small signed desktop app. There is no Docker and no separate VPN tooling to set up.
- 3Open the .aiemp with the Runner. A static dashboard opens immediately; a live (Streamlit) dashboard provisions Python on first run (about a minute) and then launches.
Pasting a connection URL
When a live dashboard needs a database, the Runner shows a "Connection details" dialog. Paste your connection string there. It is stored only in your operating-system keychain (macOS Keychain, Windows Credential Manager) — never inside the bundle and never in the cloud — and is passed to the dashboard as an environment variable when it starts.
PostgreSQL: postgresql://user:password@host:5432/dbname
MySQL: mysql://user:password@host:3306/dbname
SQLite, which is a single local file: sqlite:///relative.db, or sqlite:////absolute/path/to/file.db with four slashes for an absolute path. A plain file path works too.
Use the show/hide toggle to verify what you pasted. The "Test connection" button checks that a networked database (PostgreSQL or MySQL) is reachable before starting — handy for catching a VPN that is not connected. SQLite is a local file, so no test is required; just save and run.
Static vs live dashboards
A static dashboard is a self-contained HTML report with its data baked in. It opens instantly and needs no configuration.
A live dashboard (Streamlit) runs a small Python server on your machine and queries your database in real time. The Runner manages its own Python through uv, so you install nothing beyond the Runner itself.
What stays private
The dashboard is served on 127.0.0.1, your machine only, and is never reachable from the network. Connection secrets stay in your OS keychain. Closing the dashboard window stops the local server.