Codex Persistent MCP
C

Codex Persistent MCP

An MCP server that implements session persistence through the local codex-cli, enabling MCP calls to create real recoverable sessions and supporting users to continue the conversation later using codex resume.
2 points
5.1K

What is Codex Persistent MCP?

This is a Model Context Protocol (MCP) server specifically designed for the Codex CLI, providing session persistence functionality. When an AI assistant (such as Claude, Antigravity, etc.) invokes Codex via MCP, it creates real persistent sessions that are saved locally. Users can later use the codex resume command at any time to continue the previous conversation.

How to use Codex Persistent MCP?

You can add Codex Persistent MCP to your AI assistants (such as Claude Code, Antigravity, etc.) through simple commands. After adding, the AI assistant can invoke Codex via the MCP protocol, and all conversations will be automatically saved as recoverable sessions.

Use cases

Suitable for scenarios that require long-term tracking of development conversations, code reviews, project planning, etc. Particularly suitable for development teams, technical documentation writing, code quality reviews, and other workflows that require continuous conversations and context preservation.

Main features

Session persistence
All conversations via MCP will create real persistent sessions, which are saved in the session storage of the Codex CLI and support recovery at any time.
Seamless integration
Easily integrate into AI assistants that support MCP, such as Claude Code and Antigravity, without complex configuration.
Session recovery
Users can use the codex resume <session_id> command to recover any session at any time and continue the previous conversation.
Context preservation
The context will not be lost when the server restarts. All contexts are saved in the session storage of the Codex CLI.
Multi-tool support
Provides three main tools: codex_chat, codex_plan, and codex_review to meet different usage scenarios.
Working directory awareness
Automatically identify and manage the working directories of different projects to ensure that sessions are associated with the correct workspaces.
Advantages
True session persistence: All MCP calls create real recoverable sessions
Seamless user experience: Users can continue the conversation through the familiar codex resume command
Context preservation: The conversation history will not be lost when the server restarts
Flexible integration: Supports multiple AI assistants and development environments
Project awareness: Automatically associate sessions with working directories for easy management
Limitations
Depends on Codex CLI: Requires local installation and configuration of Codex CLI
Path configuration: Manual setting of the CODEX_BIN environment variable is required in some environments
Node.js requirement: Requires a Node.js 18+ runtime environment
Session serialization: Requests for the same session need to be processed serially, which may affect concurrent performance

How to use

Install Codex CLI
Ensure that your system has installed Codex CLI (recommended version 0.77.0 or higher).
Add the MCP server to the AI assistant
Select the corresponding command to add the Codex Persistent MCP server according to the AI assistant you are using.
Verify the installation
Verify that the MCP server has been successfully added and is running normally.
Start using
Now your AI assistant can invoke Codex via MCP, and all conversations will be automatically persisted.

Usage examples

Code review workflow
Before submitting code, invoke Codex via the AI assistant for code quality review and create a traceable review session.
Project planning session
Collaborate with the AI assistant to formulate a project development plan, and all discussions are saved as persistent sessions.
Technical question consultation
Consult Codex on complex technical issues and create a long-term technical conversation record.

Frequently Asked Questions

Why do I need to set the CODEX_BIN environment variable?
How can I find the absolute path of codex?
Why do I need to pass the cwd (working directory) parameter?
How can I recover a session created by MCP?
What is the difference between a session created by MCP and a normal Codex session?
Can I override the default model and inference settings?

Related resources

Codex CLI official documentation
Complete usage documentation and API reference for Codex CLI
Model Context Protocol specification
Official specification and implementation guide for the MCP protocol
Codex Persistent MCP source code
GitHub repository for this project, containing the complete source code
Claude Code MCP guide
Detailed guide on how to use the MCP server in Claude Code

Installation

Copy the following command to your Client for configuration
Note: Your key is sensitive information, do not share it with anyone.

Alternatives

A
Airweave
Airweave is an open - source context retrieval layer for AI agents and RAG systems. It connects and synchronizes data from various applications, tools, and databases, and provides relevant, real - time, multi - source contextual information to AI agents through a unified search interface.
Python
7.1K
5 points
V
Vestige
Vestige is an AI memory engine based on cognitive science. By implementing 29 neuroscience modules such as prediction error gating, FSRS - 6 spaced repetition, and memory dreaming, it provides long - term memory capabilities for AI. It includes a 3D visualization dashboard and 21 MCP tools, runs completely locally, and does not require the cloud.
Rust
4.6K
4.5 points
M
Moltbrain
MoltBrain is a long-term memory layer plugin designed for OpenClaw, MoltBook, and Claude Code, capable of automatically learning and recalling project context, providing intelligent search, observation recording, analysis statistics, and persistent storage functions.
TypeScript
5.2K
4.5 points
B
Bm.md
A feature-rich Markdown typesetting tool that supports multiple style themes and platform adaptation, providing real-time editing preview, image export, and API integration capabilities
TypeScript
3.8K
5 points
S
Security Detections MCP
Security Detections MCP is a server based on the Model Context Protocol that allows LLMs to query a unified security detection rule database covering Sigma, Splunk ESCU, Elastic, and KQL formats. The latest version 3.0 is upgraded to an autonomous detection engineering platform that can automatically extract TTPs from threat intelligence, analyze coverage gaps, generate SIEM-native format detection rules, run tests, and verify. The project includes over 71 tools, 11 pre-built workflow prompts, and a knowledge graph system, supporting multiple SIEM platforms.
TypeScript
5.1K
4 points
P
Paperbanana
Python
7.3K
5 points
B
Better Icons
An MCP server and CLI tool that provides search and retrieval of over 200,000 icons, supports more than 150 icon libraries, and helps AI assistants and developers quickly obtain and use icons.
TypeScript
5.9K
4.5 points
A
Assistant Ui
assistant - ui is an open - source TypeScript/React library for quickly building production - grade AI chat interfaces, providing composable UI components, streaming responses, accessibility, etc., and supporting multiple AI backends and models.
TypeScript
6.6K
5 points
N
Notion Api MCP
Certified
A Python-based MCP Server that provides advanced to-do list management and content organization functions through the Notion API, enabling seamless integration between AI models and Notion.
Python
20.4K
4.5 points
G
Gitlab MCP Server
Certified
The GitLab MCP server is a project based on the Model Context Protocol that provides a comprehensive toolset for interacting with GitLab accounts, including code review, merge request management, CI/CD configuration, and other functions.
TypeScript
24.6K
4.3 points
M
Markdownify MCP
Markdownify is a multi-functional file conversion service that supports converting multiple formats such as PDFs, images, audio, and web page content into Markdown format.
TypeScript
35.5K
5 points
D
Duckduckgo MCP Server
Certified
The DuckDuckGo Search MCP Server provides web search and content scraping services for LLMs such as Claude.
Python
73.1K
4.3 points
U
Unity
Certified
UnityMCP is a Unity editor plugin that implements the Model Context Protocol (MCP), providing seamless integration between Unity and AI assistants, including real - time state monitoring, remote command execution, and log functions.
C#
31.2K
5 points
F
Figma Context MCP
Framelink Figma MCP Server is a server that provides access to Figma design data for AI programming tools (such as Cursor). By simplifying the Figma API response, it helps AI more accurately achieve one - click conversion from design to code.
TypeScript
65.6K
4.5 points
G
Gmail MCP Server
A Gmail automatic authentication MCP server designed for Claude Desktop, supporting Gmail management through natural language interaction, including complete functions such as sending emails, label management, and batch operations.
TypeScript
22.1K
4.5 points
M
Minimax MCP Server
The MiniMax Model Context Protocol (MCP) is an official server that supports interaction with powerful text-to-speech, video/image generation APIs, and is suitable for various client tools such as Claude Desktop and Cursor.
Python
48.0K
4.8 points
AIBase
Zhiqi Future, Your AI Solution Think Tank
© 2026AIBase