cortexdb
Local-first AI agent memory with a knowledge graph, ontology constraints, and traceable reasoning—all in one SQLite file.
Automatically checked · GitHub checked 2026-11-16

Overview
CortexDB
Memory your coding agents share. Tell Claude Code something once, and Codex, your other machines and next week's session already know it.

Install
Claude Code:
/plugin marketplace add liliang-cn/cortexdb
/plugin install cortexdb@cortexdb
Codex:
codex plugin marketplace add liliang-cn/cortexdb && codex plugin add cortexdb@cortexdb
Go:
go get github.com/liliang-cn/cortexdb/v2
Also for DeepSeek Harness, Obsidian, and Rust, Python and Node clients (cortexdb-client).
What you get
- Agents remember. Preferences, decisions and project facts are captured as you work and recalled before the next prompt.
- One brain, many agents. Point every agent and machine at one
cortexdb-grpcserver and they share it. - See what it knows. A live web view of the graph, every recall and every change, on desktop or phone.
- No model required. Works without an embedding model; add one for semantic search.
- A real knowledge graph. SPARQL, Cypher, ontology, inference and history, if you want to go deeper.
In Go
db, _ := cortexdb.Open(cortexdb.DefaultConfig("brain.db"))
brain := db.KnowledgeMemory()
brain.Remember(ctx, cortexdb.KnowledgeMemoryRememberRequest{MemoryID: "tabs", UserID: "alice", Content: "Alice prefers tabs."})
rec, _ := brain.Recall(ctx, cortexdb.KnowledgeMemoryRecallRequest{Query: "what does Alice prefer?", UserID: "alice"})
fmt.Println(rec.ContextPack.Text)
Docs
Guide · Plugin · Live view · Claude Code mod · Environment variables · Integrating your agent · Examples · Changelog · 中文
MIT