Update an existing endpoint connection. See Endpoint Connections.
API key authorization. Pass your API key in the Authorization header as a Bearer token. Both new (gsk_*) and legacy (gsk-) API keys are accepted, e.g. Authorization: Bearer gsk_... or Authorization: Bearer gsk-....
EndpointConnection ID
"prod_123"
"My API Connection"
"https://api.example.com/v1/chat/completions"
GET, POST, PUT, PATCH, DELETE "POST"
NONE, BEARER, API_KEY, BASIC, OAUTH2 "BEARER"
Jinja2/Nunjucks template for request body. Must contain {{ input.user_message }} placeholder.
"{\"prompt\": \"{{ input.user_message }}\"}"
JSONPath expressions to extract values from API response. Must include "output" key.
{ "output": "$.choices[0].message.content" }"ec_123"
"user_123"
Whether an auth token is configured (token value is not exposed)
Username for Basic auth
Whether a password is configured (password value is not exposed)
{ "Content-Type": "application/json" }Optional JSON Schema (draft 2020-12) defining the structure of input fields, their types, and valid values. Must have type "object" with a "properties" key.
{
"type": "object",
"properties": {
"user_message": {
"type": "string",
"description": "The main user message to the agent"
},
"reasoning_level": {
"type": "string",
"enum": ["low", "medium", "high"],
"description": "How much reasoning the agent should apply"
},
"topic": {
"type": "string",
"enum": ["billing", "shipping", "returns", "account"],
"description": "Conversation topic"
},
"customer_id": {
"type": "string",
"pattern": "^C-\\d{3,6}$",
"description": "Customer identifier"
}
},
"required": ["user_message"]
}INITIALIZATION, CONVERSATION, FINALIZATION "CONVERSATION"
Timeout in seconds
60
Maximum requests per minute
60
Whether automatic retry is enabled for failed requests
Maximum number of retry attempts (applied when retryEnabled is true)
0 <= x <= 10Initial delay in milliseconds before first retry
x >= 0Strategy for increasing delay between retries
FIXED, EXPONENTIAL, LINEAR Maximum delay cap in milliseconds
x >= 0HTTP status codes that should trigger a retry
100 <= x <= 599Endpoint connection updated successfully
"prod_123"
"My API Connection"
"https://api.example.com/v1/chat/completions"
GET, POST, PUT, PATCH, DELETE "POST"
NONE, BEARER, API_KEY, BASIC, OAUTH2 "BEARER"
Jinja2/Nunjucks template for request body. Must contain {{ input.user_message }} placeholder.
"{\"prompt\": \"{{ input.user_message }}\"}"
JSONPath expressions to extract values from API response. Must include "output" key.
{ "output": "$.choices[0].message.content" }"ec_123"
"user_123"
Whether an auth token is configured (token value is not exposed)
Username for Basic auth
Whether a password is configured (password value is not exposed)
{ "Content-Type": "application/json" }Optional JSON Schema (draft 2020-12) defining the structure of input fields, their types, and valid values. Must have type "object" with a "properties" key.
{
"type": "object",
"properties": {
"user_message": {
"type": "string",
"description": "The main user message to the agent"
},
"reasoning_level": {
"type": "string",
"enum": ["low", "medium", "high"],
"description": "How much reasoning the agent should apply"
},
"topic": {
"type": "string",
"enum": ["billing", "shipping", "returns", "account"],
"description": "Conversation topic"
},
"customer_id": {
"type": "string",
"pattern": "^C-\\d{3,6}$",
"description": "Customer identifier"
}
},
"required": ["user_message"]
}INITIALIZATION, CONVERSATION, FINALIZATION "CONVERSATION"
Timeout in seconds
60
Maximum requests per minute
60
Whether automatic retry is enabled for failed requests
Maximum number of retry attempts (applied when retryEnabled is true)
0 <= x <= 10Initial delay in milliseconds before first retry
x >= 0Strategy for increasing delay between retries
FIXED, EXPONENTIAL, LINEAR Maximum delay cap in milliseconds
x >= 0HTTP status codes that should trigger a retry
100 <= x <= 599