Hindsight
by Vectorize
Gives agents a self-hosted long-term memory server that retains facts, recalls them with parallel semantic, keyword, graph, and time search, and reflects over them, via SDKs, REST, and MCP.
Skills
Retain Memories
Stores new information in a memory bank, using an LLM to extract facts, entities, relationships, and timestamps into search indexes.
Recall Memories
Searches a bank with semantic, BM25 keyword, graph, and temporal retrieval in parallel, then fuses and reranks results to a token limit.
Reflect Over Memories
Analyzes existing memories to answer questions that need reasoning rather than lookup, shaped by the bank's disposition traits.
Mental Models
Keeps standing answers and wiki-style knowledge pages for a bank that Hindsight rewrites in the background as the bank learns more.
Coding Agent Memory
Installs per-repo memory into Claude Code, Codex, Cursor, and other CLI agents, built from git history and past sessions automatically.
Related Agents
GBrain
Gives existing agent harnesses a self-hosted memory and skills layer with sourced facts, hybrid search, and cited synth…
LightRAG
Builds a graph-based retrieval index over documents so agents can answer questions spanning many sources, served throug…
Cognee
Builds queryable memory for agents by turning documents and conversations into a knowledge graph plus vector store, sel…
LanceDB
Stores and searches multimodal vectors in an embedded database built on the Lance columnar format, running in-process o…