Install, connect Ollama, start chatting.

1. Install Ollama

Vashti talks to Ollama backends, so install Ollama first if it is not already running.

Get Ollama from ollama.com/download.

2. Pull a model

Pull at least one model in Ollama before testing chat. For example:

3. Install Vashti

Run the install command on the Linux machine that should host Vashti:

The installer downloads the latest release, verifies the checksum, installs the binary, creates a systemd service, and prints the browser URLs to try.

4. Open Vashti

After install, open one of the URLs printed by the installer. Use the localhost URL on the server itself, or the network URL from another device on the same LAN or VPN.

5. Create the first account

The first account created becomes the admin account. After an admin exists, new accounts start disabled until an admin approves them.

6. Connect models

Open Settings, check the Backends and Models sections, and confirm your Ollama models are visible. If localhost detection does not find Ollama, add the backend URL manually.

7. Update later

To update Vashti, run the same install command again. The installer replaces the binary and restarts the service without deleting your data.