AI 记忆功能简直“离谱”到快要违法了
10 个开源 GitHub 项目，让 AI Agent 无需在每次新会话时都从零开始
01 Mem0
▸ https://github.com/mem0ai/mem0
→ 66K+ stars
02 Hindsight
▸ https://github.com/vectorize-io/hindsight
→ retain → recall → reflect
03 memU
▸ https://github.com/NevaMind-AI/memU
TURN MEMORY INTO KNOWLEDGE
04 Cognee
▸ https://github.com/topoteretes/cognee
→ documents + code + conversations → knowledge graph
05 Graphiti
▸ https://github.com/getzep/graphiti
→ remembers how facts change over time
06 OpenViking
▸ https://github.com/volcengine/OpenViking
MAKE THE AGENT STATEFUL
07 Letta
▸ https://github.com/letta-ai/letta
→ memory + identity across sessions
08 Letta Code
▸ https://github.com/letta-ai/letta-code
REMEMBER ACROSS THE STACK
09 OpenMemory
▸ https://github.com/mem0ai/openmemory
10 Agent Memory Benchmark
▸ https://github.com/vectorize-io/agent-memory-benchmark
the loop:
experience → remember → connect → retrieve → act → update
3 builds I'd actually test:
coding:
Hindsight → Cognee → Letta Code
personal agent:
Mem0 → Graphiti → Letta
company brain:
Cognee → Graphiti → Hindsight
and this is where the sh...t gets interesting:
bigger context isn't the same as memory
a model can hold 1M tokens and still meet you for the first time every morning
these projects change that
save this before your agent forgets it ⭣
https://x.com/Lummox_eth/status/2104218875810185287/video/1