• Picks
  • Discover
  • Alternatives
  • About

© 2026 Willet

  • Newsletter
  • About
  • FAQ
  • Terms
  • Privacy & Cookies
Join the Discord (opens in a new tab)
vectorize-io avatar

Hindsight

Long-term memory server for AI agents that builds knowledge over time

  • Agent frameworks and tools
  • AI models and runtimes
View on GitHub
Website

What it is

Hindsight is a memory service for AI agents, run as a server with a web UI or embedded in Python. Agents store facts with retain, search them with recall, and get answers grounded in accumulated observations and mental models with reflect, all organized into separate memory banks. Developers building agents and LLM apps connect through Python, TypeScript, Go, or CLI clients, an MCP server, or framework integrations, using any of 25+ hosted or local model providers.

Images

Facts

Stars
51,927 GitHub stars.
51,927

Forks
6,179
Commits
3,766
Contributors
288
Language
Python
License
MIT
Last commit
Oct 20, 2026
On GitHub since
October 2025
First release
December 2025
Platforms
web, CLI, self-hosted, library
Self-hostable
Yes
Alternative to
Zep
Tags
  • mcp-server
  • semantic-search
  • ai-memory
  • ai-agents
  • knowledge-graph

Last synced from GitHub on Oct 20, 2026

The 10 best overlooked open-source projects, every Monday.

Share and more

Comments

No comments yet. Start the conversation.

Similar projects

  • Topoteretes avatar

    Cognee

    Knowledge-graph memory layer that gives AI agents long-term recall

    32.6k
  • Memvid avatar

    Memvid

    Single-file memory store for AI agents with built-in search

    16.6k
  • rUv avatar

    RuVector

    Rust vector database with graph memory and local embeddings for agents

    4.6k