VEvectorize
Memory

Dry-run fact extraction (preview, no persistence)

Changed on

Preview what the retain step would extract from text WITHOUT changing the bank — no entity resolution, links, embeddings, or persistence. Returns the candidate facts and the LLM token usage. Every prompt-affecting setting (retain mission, extraction mode, chunk size, …) is overridable in the body to A/B a candidate config against the bank's current one. This is a read-only tool: nothing is stored.

post/v1/default/banks/{bank_id}/memories/dry-run-extract

Request

Path parameters

bank_idstring required

Headers

authorizationstring nullable

Request body

contextstring

Optional context about the content.

timestampstring date-time nullable

Reference timestamp for resolving relative times (ISO 8601).

agent_namestring nullable

Deprecated: describe the speaker in context instead. Narrator override (memory owner) primed in the prompt; still honored for backwards compatibility.

strategystring nullable

Name of a retain strategy to extract under (a key of the bank's retain_strategies). Omit it and the bank's retain_default_strategy applies, exactly as it does for a retain that names none.

retain_missionstring nullable
retain_extraction_modestring nullable
retain_custom_instructionsstring nullable
retain_extract_causal_linksboolean nullable
retain_chunk_sizeinteger nullable
entities_allow_free_formboolean nullable
llm_output_languagestring nullable

Response

Successful Response

Example response

{
  "usage": {
    "input_tokens": 1500,
    "output_tokens": 500,
    "total_tokens": 2000
  }
}

Changes