Skip to main content
PATCH
/
api
/
v1
/
agent
/
{id}
/
config
Update Agent
curl --request PATCH \
  --url https://api.example.com/api/v1/agent/{id}/config

​
Path Parameters

ParameterTypeDescription
idstringThe agent’s unique identifier

​
Request Body

Only include fields you want to change: 
{
  "system_prompt": "You are a helpful assistant.",
  "llm_config": {
    "temperature": 0.5
  },
  "enable_planning": true,
  "enable_reflection": false,
  "knowledge_bases": ["550e8400-e29b-41d4-a716-446655440000"]
}
Capability toggles are mutable on this endpoint. Request-time overrides can still enable or disable planning and reflection for a single run without changing the stored agent config. Assigned knowledge bases are searched when a chat request passes knowledge_bases. Retrieval can also be invoked directly through the capability endpoint.

​
Response

Returns the updated AgentConfigResponse. 
{
  "name": "my-assistant",
  "description": "General-purpose assistant",
  "icon": "sparkles",
  "llm_config": {
    "model": "openai/gpt-4o",
    "temperature": 0.5,
    "max_tokens": 8192
  },
  "system_prompt": "You are a helpful assistant.",
  "custom_instructions": null,
  "additional_context": null,
  "max_turns": 10,
  "enable_planning": true,
  "enable_reflection": false,
  "max_reflection_attempts": 1,
  "knowledge_bases": ["550e8400-e29b-41d4-a716-446655440000"],
  "tool_count": 1,
  "sub_agent_count": 0
}