Chat

Conversational AI through Gommo's platform chat API on https://api.gommo.net.

Endpoint (recommended)

POST https://api.gommo.net/api/v2/chat
Authorization: Bearer {access_token}
Content-Type: application/x-www-form-urlencoded

action=chat&domain=%REGISTRATION_DOMAIN%&query=Hello&messages=[{"role":"user","text":"Hello"}]

Form may use access_token instead of Bearer header — same token.

Actions

action	Behavior
chat	Single JSON response
stream	SSE stream
set_model	Change chat model for session
agent	set_model (best-effort) then chat — text agent flow
models	List available chat models

Non-empty messages

Upstream action=chat requires messages with at least one entry — e.g. { "role": "user", "text": "..." }.

Streaming

Set action=stream for token-by-token SSE. Consume the stream on the client — do not expect a single JSON body.

Credits

Chat consumes Gommo user credits. Check balance via POST https://api.gommo.net/ai/me.

Portal chat

/app/chat/ uses the same upstream API. Chat history in the portal is local-only — export/import JSON from the sidebar for backup.

Playground (REST test)

When you run Gommo Developer locally (npm run docs:stack), Playground may call same-origin /gateway/chat (dev proxy only). Production integrators use POST https://api.gommo.net/api/v2/chat — not /gateway/*.

Playground — Platform → Chat worker (?worker=chat) for interactive tests. Optional JSON language (VI, EN). Full UI: /app/chat/.

Optional: self-host (operators)

When self-hosting this repo — Deploy → Self-host (advanced).

Full API

→ Chat reference — includes models, SSE parsing, chat sessions (/api/v2/ai-chat-sessions), optional form fields · Gommo public API

Next

→ Audio & TTS · Media jobs