mnemo.verify()

Check whether a vector matches a trusted source registered in your workspace. Returns a verification result with status, confidence, and source information. Metered per call.

SDK signature

mnemo.verify(
  vector: list[float],
  expected_source: str | None = None,
) -> VerifyResult

HTTP request

POST /v1/verify
Authorization: Bearer <api-key>
Content-Type: application/json

{
  "vector": [0.0, 0.1, ..., -0.04]
}

Response

{
  "verified": true,
  "confidence": 0.99,
  "source_uid": "doc:pinecone:tour-index/9c4d",
  "status": "verified"
}

Returns

Field	Type	Description
verified	bool	True if the vector matches a trusted source in the workspace.
confidence	float	0.0 to 1.0.
source_uid	str \| null	Source UID this vector was registered under, or null on no match.
status	enum	verified · rejected · modified.

Core API · Reference

mnemo.embed()

Register a vector with Mnemo before you store it. Returns the registered vector (functionally equivalent to your input — cosine similarity > 0.99) and a source UID you can use later in verify().

SDK signature

mnemo.embed(
  vector: list[float],
  source: str,
) -> EmbedResult

HTTP request

POST /v1/embed
Authorization: Bearer <api-key>
Content-Type: application/json

{
  "vector": [0.0, 0.1, ..., -0.04],
  "source": "docs/policy.pdf"
}

Response (example)

{
  "source_uid": "doc:pinecone:tour-index/9c4d",
  "vector": [0.0001, 0.1004, ..., -0.0398],
  "dimensions": 768,
  "created_at": "2026-05-07T18:30:42.182Z"
}

Field shapes shown are illustrative. Use the SDK return type as the source of truth.

Returns

Field	Type	Description
source_uid	str	Workspace-scoped UID for this vector. Use it in your store and pass it to verify().
vector	list[float]	Functionally equivalent to your input — cosine similarity > 0.99.
dimensions	int	Length of the returned vector.
created_at	iso-timestamp	Registration time.

Minimum vector dimension is 128. 768+ recommended.

Full reference docs are in progress. For pilot access and integration help, request a pilot.