docs/integrations/codex.mdx
Add persistent memory to OpenAI Codex with the Mem0 plugin. Codex forgets everything between tasks. This plugin fixes that by connecting to Mem0's cloud memory layer via MCP, automatically capturing learnings at key lifecycle points, and retrieving relevant context before every response.
Before setting up Mem0 with Codex, ensure you have:
A Mem0 Platform account and API key:
m0-)OpenAI Codex access
Your API key added to your shell profile (persists across sessions):
echo 'export MEM0_API_KEY="m0-your-api-key"' >> ~/.bashrc
source ~/.bashrc
Install the full plugin including MCP server, lifecycle hooks, and SDK skill.
Add the Mem0 marketplace:
codex plugin marketplace add mem0ai/mem0
Install the plugin:
codex plugin add mem0@mem0-plugins
Or, in the app: restart Codex, open the Plugin Directory, browse the Mem0 Plugins marketplace, and install Mem0.
The fastest way to connect Codex to Mem0 needs no plugin or marketplace. Add the MCP server with a single command:
codex mcp add mem0 --url https://mcp.mem0.ai/mcp/ --bearer-token-env-var MEM0_API_KEY
Or add it manually to ~/.codex/config.toml:
[mcp_servers.mem0]
url = "https://mcp.mem0.ai/mcp/"
bearer_token_env_var = "MEM0_API_KEY"
Make sure MEM0_API_KEY is exported in the shell you launch Codex from, then restart Codex.
This gives you the MCP tools but not the lifecycle hooks or SDK skill.
codex plugin marketplace upgrade # pull latest plugin versions
codex plugin remove mem0@mem0-plugins # uninstall the plugin (keeps the marketplace)
codex plugin marketplace remove mem0-plugins # unregister the marketplace entirely
To update, run codex plugin marketplace upgrade to pull the latest from the Mem0 repo.
Codex Cloud tasks run setup scripts and the agent in separate phases with different variable scoping:
Because the mem0 MCP server authenticates on every tool call the agent makes (not just during setup), set MEM0_API_KEY as an Environment Variable in your Codex Cloud environment configuration, not as a Secret. A Secret will let a setup script authenticate but the agent will lose access to MEM0_API_KEY once the task phase begins, breaking Mem0 MCP calls.
Lifecycle hooks that shell out to local scripts (Option A) are not applicable in Codex Cloud's ephemeral containers; use Option B (Direct MCP) with MEM0_API_KEY set as above.
| Component | Plugin Install | MCP Only |
|---|---|---|
| MCP Server (9 memory tools) | Yes | Yes |
| Lifecycle Hooks | Opt-in (see below) | No |
| Mem0 SDK Skill | Yes | No |
Once installed, the following tools are available in every Codex session:
| Tool | Description |
|---|---|
add_memory | Save text or conversation history for a user/agent |
search_memories | Semantic search across memories with filters |
get_memories | List memories with filters and pagination |
get_memory | Retrieve a specific memory by ID |
update_memory | Overwrite a memory's text by ID |
delete_memory | Delete a single memory by ID |
delete_all_memories | Bulk delete all memories in scope |
delete_entities | Delete a user/agent/app/run entity and its memories |
list_entities | List users/agents/apps/runs stored in Mem0 |
Unlike Claude Code, Codex has no plugin-host mechanism for auto-wiring hooks from an installed plugin: it only reads hooks from ~/.codex/hooks.json (or <repo>/.codex/hooks.json). Installing the plugin (Option A) does not turn hooks on by itself. To enable them, run the bundled installer once against your local clone:
python3 <path-to-your-clone>/integrations/mem0-plugin/scripts/install_codex_hooks.py
This merges Mem0's entries into ~/.codex/hooks.json and is idempotent (safe to re-run after upgrading). It also requires the codex_hooks feature flag in ~/.codex/config.toml:
[features]
codex_hooks = true
The installer prints a reminder if the flag isn't set. Restart Codex after installing hooks or editing the config. To remove: python3 .../install_codex_hooks.py --uninstall.
Once enabled, Mem0 hooks into Codex's lifecycle to automatically manage memory:
| Hook | Event | What it does |
|---|---|---|
| Session start | SessionStart | Loads prior memories and displays status banner |
| User prompt | UserPromptSubmit | Searches relevant memories before each message |
| Pre-tool (3 handlers) | PreToolUse | Blocks MEMORY.md writes; enforces user_id/app_id on mem0 tool calls; scans files being read for relevant memory context |
| Post-tool | PostToolUse | Tracks stats, scans bash errors for related memories |
| Stop | Stop | Stores a session summary at the end of every assistant turn (not just at session end) |
| Pre-compact | PreCompact | Stores a summary before the context is compacted |
What you type is stored as yours. What Codex produces (session summaries and compaction summaries) is stored as the assistant's, so its suggestions never become your stated preferences.
# Task 1: Setting up a new service
You: Create a REST API for the notifications service using Express and TypeScript.
# Codex searches memories, finds your preferences from prior tasks.
# Mem0 stores what you said as yours:
# - Your preference: "Prefers explicit error types over generic catch-all"
# ...and what Codex did as the assistant's, in the session summary:
# - Decision: "Notifications service uses Express + TypeScript + Zod validation"
# - Convention: "All API routes follow /api/v1/{resource} pattern"
# Task 2 (days later): Extending the service
You: Add WebSocket support for real-time notification delivery.
# Codex searches memories, retrieves the architecture decisions and conventions.
# Follows the same patterns established in the first task.
MEM0_API_KEY is set: echo $MEM0_API_KEYmem0 MCP / "tool collision" errors: You combined Option A with Option B. Remove the [mcp_servers.mem0] block from ~/.codex/config.toml; the plugin registers it automaticallyscripts/install_codex_hooks.py (see Lifecycle Hooks), confirm codex_hooks = true is set under [features] in ~/.codex/config.toml, and restart Codex. MCP-only installs (Option B) never include hooks