Cognee
codexLong-term memory for Codex: Cognee captures your sessions into a knowledge graph and recalls relevant context on every prompt, with knowledge-graph and codebase skills.
10 skills
Persistent memory for Claude Code, built on Cognee's knowledge graph. Captures your prompts, tool calls and answers, recalls what's relevant on every prompt, and carries decisions and lessons into later sessions. Runs locally or on Cognee Cloud.
Plugin cognee-memory from marketplace cognee in topoteretes/cognee-integrations. Commands read from the repo's manifest — yesterday.
This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.
cognee-architectureExplain a repository's architecture (modules, routes, storage, services), find risky hotspots, draw a Mermaid diagram of the code graph, or review how far a change reaches, from Cognee's code graph. E.g. "explain the architecture of this repo", "how is this codebase structured?", "where are the hots
cognee-codeIndex a repository (local path or git URL) into Cognee's code graph and answer exact structural queries with no LLM calls: callers, dependencies, impact of changing a symbol, call paths, endpoints, re-index changes. E.g. "index this repo", "what calls or uses X?", "what breaks if I change X?", "how
cognee-digestA digest of a week or more of the user's own coding-agent sessions (default 7 days): sessions per day and project, most-edited files and the lessons Cognee learned, written up as shipped, decided and still open. E.g. "weekly digest", "recap my week", "prep my retro", "what did I get done this week?"
cognee-forgetForget something stored in Cognee memory: find the documents in the current dataset whose content matches by meaning, confirm, then delete them and their graph knowledge. Deletes a whole dataset only on explicit request. E.g. "forget what we talked about X", "delete that from memory", "remove what y
cognee-handoffWrite a concise handoff note for the current work (goal, decisions, open items, next steps, branch and PRs) and save it in Cognee so a later session can pick up where the work stopped. When pausing or wrapping up, e.g. "write a handoff", "I'm done for today, save progress", "leave notes for the next
cognee-how-it-worksExplain what the Cognee plugin captures, recalls, stores and sends, where its data and logs live, and how to inspect or turn off each behavior; check its local or cloud connection and settings. E.g. "how does Cognee work?", "what does this plugin send to the cloud?", "does it capture secrets?", "how
cognee-keysList and revoke the per-machine API keys of this plugin's Cognee agent, so a full agent can give a new machine a key again. E.g. "list the Cognee keys", "which machines have a Cognee key?", "revoke an old machine's Cognee key", "the agent is full of keys", "free a Cognee key slot", or when a session
cognee-rememberStore a fact, preference, decision, note or file permanently in the Cognee knowledge graph, tagged as user, project or agent. Plain "remember this" requests go here, not to Claude's built-in memory. E.g. "remember that I prefer X", "save this to memory", "keep this in mind for next time", "don't for
cognee-reportEvaluate Cognee itself from local telemetry: recall hit rate, latency, saves, injected context, token overhead, a scorecard and grade, before vs after an update, and why a memory was or wasn't injected last turn. E.g. "evaluate Cognee", "how is Cognee doing?", "grade, assess or audit Cognee", "is Co
cognee-searchSearch Cognee memory beyond the automatic per-prompt recall: a deeper knowledge-graph search, more results, one category (user, project or agent), cited sources, or a one-off search of another dataset without switching. E.g. "search Cognee for X", "what do you remember about X?", "dig deeper in memo
cognee-standupA standup or status update on the user's own recent work, from the coding-agent sessions Cognee recorded (prompts, edited files, what was left open) per project. Default last 24h; also today, yesterday, 48h or a date. E.g. "prep my standup", "status update on my work", "what did I work on yesterday?
cognee-switch-datasetsSwitch this session to another Cognee dataset (saves go there, recall restricted to it), pin recall to the active one, or set it to auto. Takes a dataset name (created if new), "pin" or "auto", or lists them. Even without saying Cognee or dataset, e.g. "switch to <dataset>", "use <dataset> for this
cognee-syncSync the current session from Cognee's session cache into the permanent knowledge graph now (Q&A turns, tool feedback, distilled lessons), so it becomes searchable and shows up in digests. Also runs at session end. To get this session into long-term memory before it ends, e.g. "sync this session to
cognee-timelineA dated chronology of one topic across the user's recorded coding-agent sessions, oldest first: the lessons the graph learned about it and the prompts that named it. Default last 30 days; also all time or a date. E.g. "how did X evolve?", "when did we first discuss X?", "timeline of X", "history of
cognee-recallSearches Cognee memory (the permanent knowledge graph, and an indexed repository's code graph) to retrieve relevant context. Can filter by data category (user, project, agent). Memory is auto-recalled on every prompt; use this agent for deeper or targeted searches.
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/session-start.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/session-
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/session-context-lookup.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/store-user-prompt.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/stor
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/file-context.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/file-cont
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/store-to-session.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/store
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/store-to-session.py" --stop || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/script
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/send-unit.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/send-unit.py
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/credits-refresh.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/credit
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/credits-refresh.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/credit
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/pre-compact.py" || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/pre-compac
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_ROOT}/scripts/sync-session-to-graph.py" --session-end || python "${CLAUDE_PLUGIN_ROOT}/scripts/hook_runner.py" "${CLAUDE_PLUGIN_
Long-term memory for Codex: Cognee captures your sessions into a knowledge graph and recalls relevant context on every prompt, with knowledge-graph and codebase skills.
10 skills
Appends two or three short follow-up questions after a substantive answer to help users check key facts, probe the reasoning, and notice missing context
19 skills
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
AI-powered codebase understanding — analyze, visualize, and explain any project
9 skills · 10 agents · hooks
AgentDB (external: maintained in ruvnet/agentdb, pinned to commit 27678c3, plugin 3.6.0): agentdb's own standalone MCP server pinned to [email protected] (vector memory with get/list/update/delete for records, patterns and skills, hybrid keyword + vector episode search and an opt-in local cross-encoder, an opt-in aidefence injection guard on memory writes, reflexion, skills, outcome feedback, capabilities report), a memory-specialist agent, 3 skills, /agentdb-status, and the agentdb mod guard that keeps secrets out of memory and refuses delete/reset unless allowed. Complements ruflo-agentdb, which reaches AgentDB through ruflo's server
3 skills · 1 command · 1 agent · 1 MCP server
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
AI memory system — mine projects and conversations into a searchable palace. 47 MCP tools, auto-save hooks, guided setup.
3 skills · 6 commands · 1 MCP server · hooks
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.