{"name":"hindsight","repo":"vectorize-io/hindsight","url":"https://midorreal.com/project/vectorize-io-hindsight","repository":"https://github.com/vectorize-io/hindsight","description":"Hindsight: Agent Memory That Learns","facts":{"license":{"value":"MIT","source":"https://github.com/vectorize-io/hindsight","checked":"22 Nov 2026"},"language":{"value":"Python","source":"https://github.com/vectorize-io/hindsight","checked":"22 Nov 2026"},"homepage":{"value":"https://hindsight.vectorize.io/","source":"https://hindsight.vectorize.io/","checked":"22 Nov 2026"},"last_release":{"value":"v1.2.3, 19 Nov 2026","source":"https://github.com/vectorize-io/hindsight/releases","checked":"22 Nov 2026"},"releases":{"value":"78","source":"https://github.com/vectorize-io/hindsight/releases","checked":"20 Nov 2026"},"contributors":{"value":"291","source":"https://github.com/vectorize-io/hindsight/graphs/contributors","checked":"20 Nov 2026"},"archived":{"value":"no","source":"https://github.com/vectorize-io/hindsight","checked":"22 Nov 2026"}},"stars":62534.0,"writeup":{"what_it_is":"Hindsight is an agent memory system that enables AI agents to learn and retain information over time. It stores, recalls, and reflects on memories to improve agent performance on long-term memory tasks.","audience":"Developers building AI agents that need persistent memory","claims":[{"kind":"specific","claim":"Achieves state-of-the-art performance on the LongMemEval benchmark","excerpt":"Hindsight is the most accurate agent memory system ever tested according to benchmark performance. It has achieved state-of-the-art performance on the LongMemEval benchmark","status":"not_checked"},{"kind":"specific","claim":"Works with 25+ LLM providers","excerpt":"Hindsight works with **25+ LLM providers** via `HINDSIGHT_API_LLM_PROVIDER`","status":"not_checked"},{"kind":"specific","claim":"Offers 60+ integrations","excerpt":"**60+ integrations** — most need no code changes.","status":"not_checked"},{"kind":"specific","claim":"Can be deployed on multiple platforms including Linux, macOS, and Windows","excerpt":"| Platform | Docker | Bare Metal (pip) | Embedded DB (pg0) |\n|----------|--------|------------------|--------------------|","status":"not_checked"},{"kind":"broad","claim":"Eliminates shortcomings of RAG and knowledge graph approaches","excerpt":"It eliminates the shortcomings of alternative techniques such as RAG and knowledge graph","status":"too_broad"}],"alternatives":[{"name":"OpenViking","url":"https://midorreal.com/project/volcengine-openviking"},{"name":"ai-memory","url":"https://midorreal.com/project/akitaonrails-ai-memory"},{"name":"engram","url":"https://midorreal.com/project/gentleman-programming-engram"},{"name":"claude-mem","url":"https://midorreal.com/project/thedotmack-claude-mem"},{"name":"beads","url":"https://midorreal.com/project/gastownhall-beads"}],"written_by":"AI, from the project's README","date":"2026-10-09"},"why_now":null,"owner_supplied":null,"score":{"verdict":"real","hype":89,"reality":87,"gap":2,"momentum":0.6,"confidence":60,"version":"score-2.0","rule":"verdict-2.0","calculated_at":"2026-11-22T08:00:00+00:00","verdict_source":"rule"},"early_read":null,"changes":[{"date":"2026-11-19T22:48:28+00:00","text":"Latest release v1.2.3","source":"https://github.com/vectorize-io/hindsight/releases"},{"date":"2025-12-04T12:11:19+00:00","text":"First release","source":"https://github.com/vectorize-io/hindsight/releases"},{"date":"2025-10-30T11:49:48+00:00","text":"Repository created","source":"https://github.com/vectorize-io/hindsight"}],"cite":"hindsight is a GitHub project at https://github.com/vectorize-io/hindsight written mainly in Python, described there as \"Hindsight: Agent Memory That Learns\".","attribution":"Data from Mid or Real, https://midorreal.com/project/vectorize-io-hindsight, checked 22 November 2026. Attribution with a link is required.","still_real":null}