Skip to content
POST/api/v1/guardrails/generate

Generate guardrails from intent

Takes a natural-language safety intent (e.g., "Don't reveal customer SSNs") in `policyText` — or the SDK aliases `description` / `findings` — and uses an LLM to extract enforceable guardrail policies. Optional `category`, `action` (block/alert/redact) and `severity` set the defaults for generated rules. Returns `{guardrails, rawPolicies, confidence (0-1), skipped}`: each guardrail carries streaming-validator rules of type `keyword`, `regex` or `pii` with an action and severity (the `toxic` and `length` rule kinds exist in the validator but this generator never emits them) — regex patterns and keyword denylists, not scorer thresholds. `skipped` lists policy sentences that yielded no enforceable rule, so an inert rule is never reported as enforcement. 400 `MISSING_POLICY_TEXT` when empty, 400 `POLICY_TOO_LONG` above 50,000 characters. Requires the `firewall:check` scope on the API-key path and is rate-limited 10/min because it spends the org's BYOK OpenAI key (or the platform key). Generated rules require human review before deploy.

Authentication

Send Authorization: Bearer YOUR_API_KEY on every request. Generate API keys at /dashboard/settings/api-keys.

Request body required

Example

{
  "policyText": "string",
  "description": "string",
  "category": "string",
  "action": "string",
  "severity": "string"
}
Schema
{
  "application/json": {
    "schema": {
      "type": "object",
      "properties": {
        "policyText": {
          "type": "string",
          "maxLength": 500000
        },
        "description": {
          "type": "string",
          "maxLength": 500000
        },
        "findings": {
          "anyOf": [
            {
              "type": "string"
            },
            {
              "type": "array",
              "items": {}
            },
            {
              "type": "object",
              "additionalProperties": {}
            }
          ]
        },
        "category": {
          "type": "string",
          "maxLength": 200
        },
        "action": {
          "type": "string",
          "maxLength": 200
        },
        "severity": {
          "type": "string",
          "maxLength": 200
        }
      },
      "additionalProperties": {}
    }
  }
}

Response

200 example

{
  "success": true
}

All status codes

200Generated guardrails.
400(no description)
401(no description)
403Forbidden — the credential is valid but lacks the API-key scope, member role, or plan entitlement this operation requires.
429(no description)
500Internal Server Error — GENERATION_FAILED, LLM_KEY_MISSING.

Code samples

cURL

curl -X POST \
  https://evalguard.ai/api/v1/guardrails/generate \
  -H "Authorization: Bearer $EVALGUARD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "policyText": "string", "description": "string", "category": "string", "action": "string", "severity": "string" }'

TypeScript

// The TypeScript SDK (@evalguard/sdk) exposes TYPED methods — runEval,
// getEval, runSecurityScan, checkFirewall, … — not a generic request().
// For an arbitrary endpoint, call it directly:

const res = await fetch("https://evalguard.ai/api/v1/guardrails/generate", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.EVALGUARD_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    "policyText": "string",
    "description": "string",
    "category": "string",
    "action": "string",
    "severity": "string"
  }),
});
console.log(res.status, await res.json());

Python

# The Python SDK (pip install evalguardai) exposes TYPED methods on
# EvalGuardClient — run_eval, get_eval, … — not a generic request().
# For an arbitrary endpoint, call it directly:

import os
import requests

headers = {"Authorization": f"Bearer {os.environ['EVALGUARD_API_KEY']}"}
headers["Content-Type"] = "application/json"

response = requests.request(
    "POST",
    "https://evalguard.ai/api/v1/guardrails/generate",
    headers=headers,
    json={
    "policyText": "string",
    "description": "string",
    "category": "string",
    "action": "string",
    "severity": "string"
},
)
print(response.status_code, response.json())

Go

package main

import (
	"context"
	"fmt"
	"net/http"
	"os"
	"strings"
)

func main() {
	body := strings.NewReader(`{"policyText":"string","description":"string","category":"string","action":"string","severity":"string"}`)
	req, _ := http.NewRequestWithContext(context.Background(), "POST", "https://evalguard.ai/api/v1/guardrails/generate", body)
	req.Header.Set("Authorization", "Bearer "+os.Getenv("EVALGUARD_API_KEY"))
	req.Header.Set("Content-Type", "application/json")
	resp, err := http.DefaultClient.Do(req)
	if err != nil { panic(err) }
	defer resp.Body.Close()
	fmt.Println(resp.Status)
}

Errors

400401403429500

Other Guardrails endpoints