Chat

Chat is a signed-in playground. It talks to the ApiGo gateway — the same path your backend will use — so what you see in Chat is close to production behavior.

Before you start

1. Sign in.
2. Make sure the active workspace has balance (Workspace → Billing).
3. Optionally create a dedicated API key for experiments so playground traffic stays separate from production.

What you can do

• Start conversations and switch models from the vendor-grouped picker.
• Stream responses and retry or edit earlier turns.
• Keep local conversation history in the sidebar (including archived threads).
• Cross-check unexpected answers or latency in Workspace → Call Logs.

How it relates to the API

Chat is not a separate product API. When you are ready to automate:

1. Take the model slug you liked in Chat.
2. Call https://api.apigo.ai/v1/chat/completions with your own key.
3. Keep the same workspace so Usage and Billing stay consistent.

If Chat works but your SDK fails, compare base URL, key, and model slug first — those three mismatches cause most “works in UI, fails in code” reports.