pip install. The plugin captures your prompts, tool traces, and answers into session memory, injects relevant context on every prompt, and syncs the session into your knowledge graph on session end.
Sessions are disposable; your memory isn’t.
1. Enable hooks and install
The plugin depends on Codex lifecycle hooks. Enable them first, then install from the Codex marketplace:You can enable hooks manually instead by adding a
[features] section with hooks = true to ~/.codex/config.toml. If Codex prompts you to review the hooks, approve the Cognee hooks so it can call the plugin on prompt submit, tool use, stop, compaction, and session end.cognee: <dataset> · <mode> to confirm the plugin is active.
2. Point it at Cognee Cloud
Write both variables once into~/.cognee/.env, using your tenant base URL and an API key. The file is created with a commented template on the first session start, is shared with the Claude Code plugin, and its values act like shell exports — except you only set them once, and they survive closing the terminal:
- macOS / Linux (bash, zsh)
- Windows (PowerShell)
- Windows (cmd.exe)
codex launch.
When
COGNEE_BASE_URL is set, the plugin runs as a pure thin HTTP client to your tenant — it does not install a local Cognee runtime. When COGNEE_BASE_URL is unset, the plugin instead bootstraps a local API at http://localhost:8011. Setting the Cloud URL and key is what routes memory to your tenant.LLM_API_KEY in the same file is fine: cloud still wins, because a configured URL is what selects cloud. To send one terminal to local instead, export COGNEE_BACKEND=local before launching — unset COGNEE_BASE_URL does not work, because the env file re-injects the URL at the next launch. See Which mode wins.
3. Choose a dataset
All writes and recall are scoped to a single dataset, selected with theCOGNEE_PLUGIN_DATASET environment variable. By default both the Codex and Claude Code plugins use agent_sessions, so memory is shared across both integrations automatically. On your tenant that sharing is real rather than nominal: each plugin authenticates as its own agent sub-user, and they join one cognee-agent role holding read and write on your datasets, which the plugins then address by canonical UUID. Set COGNEE_SHARED_AGENT_MEMORY=false for separated, per-plugin memory.
Set a custom dataset at launch:
COGNEE_RECALL_DATASET_HINT=off to stop it; see Sessions & datasets for the full flow.
4. Pick a session (optional)
By default thesession_id is derived from the Codex thread, so a new conversation starts a new one and codex resume continues the same one. Set COGNEE_SESSION_ID to pin a specific named session, or to deliberately share one live session across two terminals:
5. Verify
Quit Codex — theSessionEnd hook syncs the session into Cognee (an exit-watcher fallback covers a hard exit). Then start a fresh session and ask:
What do you know from cognee?Answering from a clean session confirms it’s recalling from your Cloud memory. Alongside automatic capture, the plugin ships skills for explicit requests — ask for what you want and Codex picks the matching one:
What reaches your tenant. Prompts, tool traces, and answers are captured automatically, with credential paths (
.env, *.pem, */.ssh/*, and the rest) skipped outright and common secrets redacted before anything is sent. COGNEE_CAPTURE=false stops automatic capture entirely while leaving recall and explicit remember working — see What gets captured.
Two things about the code graph differ against a Cloud tenant. Repositories are not indexed automatically — the plugin only auto-indexes against a local server, so a private checkout is never shipped to a hosted tenant on its own initiative. Ask for it with the codebase skill and give it a git URL: your tenant clones the repo rather than reading your disk, which is also why the graph reflects your last pushed commit rather than your working tree. A local path is not an option here — the server cannot see it. See Code graph.
Memory is captured and synced from your sessions over time, and writes build the graph in the background — so a brand-new setup may recall nothing until at least one session has synced. An empty first recall is expected.
When the workspace balance runs out, every billable operation is refused with HTTP 402 and memory goes quiet. The status line says so rather than leaving you guessing, in three shapes:
credits: $0.61 · top up: https://platform.cognee.ai/billing— a dollar or less left. The link tracks the balance, so it appears before anything has actually been refused.credits: $2.04 (not enough for recall)— a request was refused while the balance still read healthy. The refused operation is named only above that one-dollar threshold; below it you get the top-up link instead.credits: not enough for recall · top up: …— refused with no balance to print, because the billing lookup itself failed.
/remember, and background improve runs each report themselves this way. The next billable operation that succeeds clears the note, so a top-up shows through on the following prompt. See When credits run out.Configuration reference
Precedence: environment variables →~/.cognee/.env → defaults. There is no config.json; older versions wrote one, and SessionStart now deletes a leftover file. The COGNEE_BACKEND / COGNEE_CODEX_BACKEND mode switch follows the same precedence. Wherever it is set, it pins that terminal’s mode regardless of where the connection variables are defined.
Update the plugin with
codex plugin marketplace upgrade cognee — the cognee marketplace tracks main, and updates are not automatic. See the full integration guide for the session-distillation and idle-watcher variables; the advanced session-sync and update-notification knobs are covered in the plugin README.
Full integration guide
Manual config, hooks reference, and debugging
Claude Code
The same memory plugin for Claude Code