Hindsight is a persistent memory system for AI agents built by Vectorize.io. It enables agents to store, search, reason over, and pre-compute reflections from memory — making them contextually aware across sessions.
Source
This documentation is sourced from docs.hindsight.vectorize.io and organized as an SOP reference site with RAG chat.
| Concept | What It Does |
|---|---|
| Memory Banks | Isolated containers for memories with profile, settings, and disposition traits |
| Retain | Store and process natural language content into categorized memories |
| Recall | Search memories using TEMPR multi-strategy retrieval (semantic, keyword, graph, temporal) |
| Reflect | AI-powered reasoning over memories with confidence scores and citations |
| Mental Models | Pre-computed cached reflections that auto-refresh as memories evolve |
# Python
pip install hindsight-client
# TypeScript / JavaScript
npm install @vectorize-io/hindsight-clienthttps://api.hindsight.vectorize.ioAuthentication: Authorization: Bearer your-api-key (keys start with hsk_)