Hindsight icon

Hindsight

Retain, recall, and reflect on long-term memory

Actions3

Reflect

AI-generated

Summary

Send a natural-language question to the specified Hindsight memory bank to receive a synthesized answer generated by a large language model (LLM) based on the bank's stored memories.

Inputs

  • Bank ID (required) — Identifier of the Hindsight memory bank to use. The bank is created automatically if it does not exist.
  • Query (required) — A natural-language question to be answered using the memories stored in the bank.
  • Budget — Controls the reflection budget level to determine the thoroughness of the LLM synthesis; options are Low, Mid (default), or High.

Output shape

a single JSON object containing the LLM-synthesized answer based on the bank's memories

The output is a JSON response from the Hindsight API's reflect endpoint, returned per input item. Errors return an error message JSON if continueOnFail is enabled.

Examples

Example 1: You want to ask a question that requires synthesizing insights from accumulated memories in a Hindsight bank, such as summarizing or reasoning about stored information.

Provide the target 'Bank ID', input the natural-language 'Query', and optionally select the 'Budget' level to control answer depth.

Links

Discussion