# Hindsight

> MIT retain→recall→reflect agent memory (+ MCP)

**Category:** AI Agents  
**Pricing:** open-source  
**URL:** https://hindsight.vectorize.io  
**GitHub:** https://github.com/vectorize-io/hindsight  
**Added:** 2026-09-25

---

## The Hook

Hindsight is Vectorize’s agent memory system built around retain → recall → reflect: agents that learn over time, not only replay conversation history — with MCP tools on the same bank.

## What's Good

MIT (vectorize-io/hindsight). Self-host via Docker, pip, or Helm; Python/TS clients; MCP server exposing retain/recall/reflect per bank. Docs, cookbook, and integrations at hindsight.vectorize.io. Cloud is optional when you do not want to run the stack yourself.

## The Catch

Vs Mem0, Graphiti, Engram, and Supermemory: Hindsight is a retain/recall/reflect memory product with optional hosted Cloud/Enterprise, not a general memory SDK alone, a temporal knowledge graph, a local binary brain, or a speed-first memory API. Treat LongMemEval / SOTA marketing as vendor claims, not catalog proof; Cloud and Enterprise are upsell paths beside the OSS self-host bar.

## Verdict

MIT agent memory when you want retain→recall→reflect (+ MCP) you can self-host; keep Cloud optional and do not bet the stack on a single eval leaderboard.