AI Agents
Protocols
Chat Completions and A2A protocols for connecting external agents.
Chat Completions
The default protocol. Compatible with OpenAI, Anthropic, Google, Groq, and any OpenAI-compatible API.
{
"api_url": "openai",
"model": "gpt-5-mini",
"protocol": "chat_completions"
}For custom endpoints:
{
"api_url": "https://your-agent.example.com/v1",
"api_key": "your-api-key",
"model": "your-model",
"protocol": "chat_completions"
}The agent-client sends the conversation history in the standard Chat Completions format with role: "user" and role: "assistant" messages.
A2A (Agent-to-Agent)
Google's A2A protocol for agent interoperability.
{
"api_url": "https://your-a2a-agent.example.com",
"protocol": "a2a"
}The agent-client sends tasks and receives responses according to the A2A specification.
Supported providers
| Provider | api_url value | Default model |
|---|---|---|
| OpenAI | "openai" | gpt-5-mini |
| Anthropic | "anthropic" | claude-sonnet-4-6 |
"google" | gemini-2.5-flash | |
| Groq | "groq" | openai/gpt-oss-20b |
| Custom | Full URL | — |
When using a built-in provider, set the provider's API key as a Supabase Edge Function secret (e.g., OPENAI_API_KEY). Or set api_key in the agent config to use a per-agent key.