Skip to main content
POST
/
v1
/
verify
Verify an answer against a tool result
curl --request POST \
  --url https://api.prompt-wall.com/v1/verify \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "prompt": "What is the capital of France?",
  "answer": "Paris is the capital of France.",
  "tool_result": "Paris",
  "verified_source_used": false,
  "grounding_required": false
}
'
{
  "ok": true,
  "answer": "<string>",
  "changed": true,
  "confidence": "high",
  "evidence_consistent": true,
  "governance": "allow",
  "mismatch_type": "none",
  "latency_ms": 123,
  "matches": [
    "<string>"
  ],
  "pipeline": {
    "scanner": "<string>",
    "policy": "<string>",
    "judge": "<string>",
    "enforcement": "<string>"
  },
  "request_id": "<string>"
}

Authorizations

​
Authorization
string
header
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json
​
prompt
string
required

The user's original question

Example:

"What is the capital of France?"

​
answer
string
required

The proposed answer to validate

Example:

"Paris is the capital of France."

​
tool_result

Source material the answer should be grounded in

Example:

"Paris"

​
verified_source_used
boolean
default:false

Whether a verified tool was consulted

​
grounding_required
boolean
default:false

Override: force grounding verification

Response

Verification result

​
ok
boolean
​
answer
string

Final answer (may be rewritten if governance=rewrite)

​
changed
boolean

Whether the answer was modified by enforcement

​
confidence
enum<string>
Available options:
high,
medium,
low
​
evidence_consistent
boolean
​
governance
enum<string>
Available options:
allow,
rewrite,
block,
regenerate
​
mismatch_type
enum<string>
Available options:
none,
contradiction,
numeric,
insufficient_evidence,
unsupported_inference
​
latency_ms
integer
​
matches
string[]

Security pattern matches (if any)

​
pipeline
object
​
request_id
string