Assistant

Updated 2026 · FAi

The assistant is in-app chat on our hosted service: streamed replies, optional tools on integrations like Notion or Discord, and model picks in chat settings. Provider and API key depth lives under Integrations.

Turn it on and configure providers

Sign in, then open Integrations → Assistant.
Use the toggles there to enable the assistant for your account, choose whether you use FAi-hosted models or your own API keys where available, and add keys for providers you use.
If your workspace uses Ollama or optional browser / web-reading features, configure those on the same page; availability depends on how your administrator set up the backend.

Open chat

From the sidebar under Work, open Assistant, or go to /app/assistant. If the assistant is not configured, the app will surface setup hints and links to Integrations.

Chat settings

In the assistant header, open Settings (gear icon).

Model — Pick a provider and model from the list loaded from the server (or common defaults if the list is unavailable). For more providers or keys, use Integrations → Assistant.
Connected apps & tools — When on, the assistant can call tools tied to integrations you have linked. Turn it off for plain Q&A without tool use.

Threads and where data lives

Conversation history for the assistant is stored in your browser on this device (for example when you create or switch threads in the sidebar). Clearing site data or using another browser or device will not show the same local history unless you use an export flow the UI provides.

Usage and billing

For token or usage summaries tied to your account, open Token Usage under Account in the sidebar. Whether usage is billed through FAi-hosted inference or your own provider keys is determined on Integrations → Assistant.