curl --request POST \
--url https://api.portkey.ai/v1/guardrails \
--header 'Content-Type: application/json' \
--header 'x-portkey-api-key: <api-key>' \
--data '
{
"name": "JWT Authentication Guard",
"organisation_id": "550e8400-e29b-41d4-a716-446655440001",
"checks": [
{
"id": "default.jwt",
"parameters": {
"jwksUri": "https://example.com/.well-known/jwks.json",
"headerKey": "Authorization",
"algorithms": [
"RS256"
],
"cacheMaxAge": 86400,
"clockTolerance": 5,
"maxTokenAge": "1d"
}
}
]
}
'{
"id": "<string>",
"slug": "<string>",
"version_id": "<string>"
}{
"error": {
"code": "<string>",
"message": "<string>",
"param": "<string>",
"type": "<string>"
}
}{
"error": {
"code": "<string>",
"message": "<string>",
"param": "<string>",
"type": "<string>"
}
}{
"error": {
"code": "<string>",
"message": "<string>",
"param": "<string>",
"type": "<string>"
}
}Create Guardrail
POST
https://api.portkey.ai/v1
/
guardrails
curl --request POST \
--url https://api.portkey.ai/v1/guardrails \
--header 'Content-Type: application/json' \
--header 'x-portkey-api-key: <api-key>' \
--data '
{
"name": "JWT Authentication Guard",
"organisation_id": "550e8400-e29b-41d4-a716-446655440001",
"checks": [
{
"id": "default.jwt",
"parameters": {
"jwksUri": "https://example.com/.well-known/jwks.json",
"headerKey": "Authorization",
"algorithms": [
"RS256"
],
"cacheMaxAge": 86400,
"clockTolerance": 5,
"maxTokenAge": "1d"
}
}
]
}
'{
"id": "<string>",
"slug": "<string>",
"version_id": "<string>"
}{
"error": {
"code": "<string>",
"message": "<string>",
"param": "<string>",
"type": "<string>"
}
}{
"error": {
"code": "<string>",
"message": "<string>",
"param": "<string>",
"type": "<string>"
}
}{
"error": {
"code": "<string>",
"message": "<string>",
"param": "<string>",
"type": "<string>"
}
}Portkey is now PRISMA AIRS AI Gateway. See it in action.
Contact Us
Authorizations
Body
application/json
Name of the guardrail
Example:
"Content Safety Check"
Target type for the guardrail. Use "llm" for LLM API requests (default) or "mcp_tools" for MCP tool calls.
Available options:
llm, mcp_tools Workspace UUID (required if organisation_id not provided and not using API key)
Organisation UUID (required if workspace_id not provided and not using API key)
Array of guardrail checks to apply. Required for "llm" target; optional for "mcp_tools" target.
Minimum array length:
1Show child attributes
Show child attributes
Actions to take when guardrail checks fail or pass. Required for "llm" target; optional for "mcp_tools" target.
Show child attributes
Show child attributes
Last modified on August 6, 2026
Was this page helpful?