Skip to content
Fast Inference
DOCS / Pi
Integration

Pi

Add Inference.net as a Pi Agent provider.

fast pi on adds an inference-net provider to Pi's models.json, stores the key in auth.json, and makes it the default in settings.json. All three files live under ~/.pi/agent.

Connect Pi

  1. Install

    bash
    npm install -g @inference/fast

  2. Activate

    bash
    fast pi on

    The first run signs you in through the browser and creates an API key for this machine. The command backs up Pi's config before editing it.

  3. Restart Pi

    Restart Pi and send a prompt; the request appears under Usage.

Choose a model

Pi uses kimi-k3-fast unless you pick another ID from the model catalog:

bash
fast pi on --model kimi-k3-fast

Run on again to change the default. Restart Pi afterward.

Switch models in Pi

Every model from the model catalog is registered under the inference-net provider, and enabledModels is scoped to it, so Ctrl+P cycles through them and /model (Ctrl+L) opens on them (Tab shows every provider). A /model choice lasts for the session unless you press Ctrl+S in the picker to save it as the default.

Check status

bash
fast pi status

Shows the connection, model, and config paths.

Disconnect

bash
fast pi off

Restores all three files to the backups taken by on, so edits made after connecting are lost. Restart Pi afterward.

Troubleshooting

  • Inference.net is not listed in Pi — restart Pi and check that models.json has an inference-net provider.
  • Requests return 401 — run fast status, then fast pi on.
  • Model rejected — pick an ID from the model catalog.