Serveur MCP Codex
Commit 14 nov. 2026Permet à un client IA de gérer un wiki personnel en markdown, d'organiser des documents dans des dossiers hiérarchiques et d'effectuer des recherches plein texte.
Aussi connu sous bocan/codex
- Étoiles
- 9
- Outils
- 16
- Téléchargements, 30 jours
- —
- Dernière version
- —
- Dernier commit
- 14 nov. 2026
- Licence
- MIT
Qu'est-ce que Codex MCP ?
Codex est un serveur MCP. Permet à un client IA de gérer un wiki personnel en markdown, d'organiser des documents dans des dossiers hiérarchiques et d'effectuer des recherches plein texte. Il expose 16 outils, comme search_pages, get_page, create_page.
Résumé généré à partir du README du projet
À propos du MCP server Codex
Codex permet aux utilisateurs individuels de gérer une base de connaissances personnelle via un wiki markdown structuré. Cette application mono-utilisateur vous permet d'organiser vos documents au sein d'une structure de dossiers hiérarchique, ce qui en fait une solution idéale pour la prise de notes personnelle et la documentation. Elle propose une interface à trois volets comprenant un arbre de dossiers, un éditeur markdown et un volet de prévisualisation en direct pour optimiser votre processus d'écriture.
Les utilisateurs peuvent interagir avec leurs données via une API RESTful ou l'interface intégrée, qui prend en charge le GitHub Flavored Markdown et les diagrammes Mermaid. Le système inclut un contrôle de version basé sur Git pour l'historique des documents et un outil de recherche plein texte pour trouver rapidement des informations. Pour sécuriser vos données privées, le serveur nécessite une connexion par mot de passe et applique une limitation de débit pour empêcher tout accès non autorisé.
L'intégration avec les modèles IA est un composant central, vous permettant de connecter une clé API Anthropic Claude ou une instance locale Ollama. Cette configuration permet à un assistant IA de fournir des réponses contextuelles basées sur vos documents ouverts, avec des réponses en streaming et des blocs de raisonnement.
Ce que vous pouvez faire avec
- 1Organiser votre documentation personnelleCréez et gérez une collection hiérarchique de fichiers markdown pour vos notes privées.
- 2Interroger votre base de connaissancesPosez des questions à un assistant IA sur le contenu spécifique de vos documents.
- 3Suivre les modifications de documentsConsultez les versions précédentes de vos notes et voyez exactement ce qui a été ajouté ou supprimé.
- 4Visualiser des flux de travail complexesGénérez et exportez des diagrammes directement dans vos pages markdown en utilisant le code Mermaid.
Extrait du README du projet. Chaque cas d'utilisation est vérifié par rapport à une citation du document.
Outils MCP Codex 16
Détectés depuis le READMEsearch_pagesget_pagecreate_pageupdate_pagedelete_pagerename_pagemove_pagelist_folderslist_pagescreate_folderdelete_folderrename_folder
| search_pages | — | |
| get_page | — | |
| create_page | — | |
| update_page | — | |
| delete_page | — | |
| rename_page | — | |
| move_page | — | |
| list_folders | — | |
| list_pages | — | |
| create_folder | — | |
| delete_folder | — | |
| rename_folder | — |
Afficher les 16 outilsAfficher moins d'outils
list_attachmentsupload_attachmentget_attachmentdelete_attachment
| list_attachments | — | |
| upload_attachment | — | |
| get_attachment | — | |
| delete_attachment | — |
Configuration
AUTH_PASSWORDMCP_API_KEYMCP_ENABLED
| Nom | Description | Type |
|---|---|---|
| AUTH_PASSWORD | — | |
| MCP_API_KEY | — | |
| MCP_ENABLED | — |
Vérifications automatisées
Exécuté le 13 nov. 2026Rapport complet →Pas encore vérifié. Un serveur nécessite un score de 80, aucune faille sérieuse et suffisamment de données pour que nous puissions effectuer les vérifications. Fonctionnement des vérifications · Ce qu'il faut corriger pour gagner des points
- Dépôt source lié et accessiblebocan/codex is public and reachablegithub.com/bocan/codex ↗Réussi
- Le nom du package correspond au dépôt sourceNon vérifié
- Provenance du build (attestation Sigstore)Non vérifié
- Licence open source (OSI)MITspdx.org/licenses/MIT.html ↗Réussi
- Commit ou version publiée au cours de l'année écouléeLast commit or release 2026-11-12github.com/bocan/codex ↗Réussi
- OpenSSF ScorecardNon vérifié
- Nom du package non proche d'un package populaireNon vérifié
- Aucun avis critique dans le package ou ses dépendancesNon vérifié
- Aucune instruction cachée destinée au modèle16 tool descriptions and the README read; no override phrases, credential paths or hidden charactersRéussi
- Les outils destructifs sont étiquetésNon vérifié
- Le point de terminaison distant répond à la poignée de main MCPNon vérifié
README
Voir sur GitHub📝 Codex
A wiki and document store built with React and Express.
Single-user personal knowledge base - A full-stack TypeScript application that provides a beautiful, intuitive interface for creating and managing markdown documents organized in a hierarchical folder structure.
⚠️ Note: Codex is designed as a single-user application. It does not support concurrent multi-user editing or collaboration features. Perfect for personal wikis, note-taking, and documentation.
✨ Features
- 📁 Folder Management: Create, delete, and rename folders in a collapsible tree view with right-click context menus
- 📝 Markdown Pages: Create and edit markdown documents with GitHub Flavored Markdown support
- 🧩 Smart Templates: Create new pages from reusable templates (stored under
data/templates/) - 📊 Mermaid Diagrams: Render Mermaid code fences in preview/reading mode, with export/download support
- 🎨 Three-Pane Layout: Folder tree (left), markdown editor (center), live preview (right)
- 📐 Fully Resizable: Drag to resize both horizontal panes (left/right) and vertical sections (folder tree/page list)
- 🌓 Theme Options: Auto-detects system theme preference with manual override - cycles through auto/light/dark/high-contrast modes
- ♿ Accessibility: Full ARIA labels, semantic HTML, high-contrast theme, and improved color contrast for WCAG compliance
- 📤 Move Pages: Elegant folder picker to move pages between folders via right-click menu
- 💾 Smart Auto-save: 10-second throttled saves with 5-second typing debounce - prevents excessive saves while keeping your work safe
- 🔄 Live Preview: Real-time markdown preview that updates instantly as you type (no waiting for saves)
- 🔗 Internal Links: Click links to other
.mdfiles to navigate within the app; anchor links scroll to headings - 📜 Version History: Git-backed version control with visual diff highlighting (green/red for additions/deletions)
- 🔍 Restore Versions: Browse and restore any previous version of your documents
- 📖 Reading Mode: Open any page in a new window for distraction-free reading
- 🔃 Synchronized Scrolling: Editor scroll position syncs to preview pane
- 📄 Auto-select README: Navigating to a folder automatically opens its README.md if present
- 🔎 Full-Text Search: Quick search across all pages with keyboard shortcut (⌘K/Ctrl+K) and relevance-ranked results
- 📑 Table of Contents: Auto-generated, collapsible TOC for easy document navigation with active section highlighting
- 🎤 Speech-to-Text: Dictate content using Web Speech API (Chrome/Edge/Safari)
- ⚡ Performance Caching: Server-side caching layer with 30-second TTL for fast folder/page loading
- 🌐 RESTful API: Programmatic access to all folder and page operations
- ✅ Tested: Comprehensive test suite for both backend and frontend
- 🎯 Collapsible Panes: Hide sidebars for distraction-free writing
- 🚀 Fast & Lightweight: Built with Vite for lightning-fast development
- 🔐 Password Protection: Simple password-based authentication to secure your data
- 🛡️ Security Features: Rate limiting, request logging, and security headers
- 🤖 AI Chat: Built-in AI assistant supporting Anthropic Claude and local Ollama models with streaming responses, thinking blocks, and document context
🧩 Smart Templates
When creating a new page, Codex can start from a template instead of a blank page.
- Templates live in
data/templates/*.md - Optional frontmatter at the top of the template file:
template: Your Template Name(display name)autoname: true|false(auto-generate a filename when creating)
- The frontmatter is stripped from the created page content automatically
📊 Mermaid Diagrams
Codex supports Mermaid diagrams in markdown via fenced code blocks:
```mermaid
graph TD
A --> B
```
- Diagrams render in the live preview and in reading mode (open in new window)
- You can download rendered diagrams as SVG
- Export flows (e.g. Word/PDF) include diagrams as images rather than raw Mermaid text
🤖 AI Chat
Codex includes an integrated AI chat assistant for asking questions about your documents:
Supported Providers
- Anthropic Claude - Cloud-based AI with optional "thinking" mode for extended reasoning
- Ollama - Run local models like Llama, Mistral, CodeLlama, and more
Features
- Streaming Responses - See AI responses as they're generated in real-time
- Document Context - AI can reference the currently open document for context-aware answers
- Thinking Blocks - Anthropic's extended thinking feature shows the AI's reasoning process (collapsible)
- Token Usage - Track input/output tokens for cost monitoring
- Chat History - Conversation persists while the app is open
- Export Chat - Download conversation history as JSON
- Custom System Prompts - Configure how the AI behaves via settings
- Multiple Accounts - Switch between different AI providers/accounts
- Markdown Rendering - AI responses render with full markdown support including code highlighting
- Copy Support - Copy individual messages or code blocks with one click
Configuration
AI accounts are configured in Settings (gear icon):
- Click the settings icon in the header
- Navigate to the "AI" section
- Add an account:
- Anthropic: Enter your API key from console.anthropic.com
- Ollama: Enter host (default: localhost) and port (default: 11434)
- Enable "AI Search" to show the AI chat button
Usage
- Click the robot icon (🤖) in the bottom-right corner to open AI chat
- Select your AI account from the dropdown
- Type your question and press Enter or click Send
- Use the thinking toggle (brain icon) for Anthropic's extended reasoning
- Click the settings icon to customize the system prompt
Ollama Setup
To use local AI models with Ollama:
# Install Ollama (macOS)
brew install ollama
# Start Ollama server
ollama serve
# Pull a model (in another terminal)
ollama pull llama3.2
ollama pull codellama # Great for code questions
Then add an Ollama account in Codex settings with host localhost and port 11434.
🔐 Security & Logging
Codex includes several security features to protect your data:
Authentication
- Password-based login with bcrypt hashing (10 salt rounds)
- Session management using httpOnly cookies (24-hour expiration)
- Authentication can be disabled by not setting
AUTH_PASSWORD(not recommended for public deployments)
Rate Limiting
- Login endpoint is rate-limited to 5 attempts per 15 minutes per IP address
- Prevents brute force password attacks
- Returns
429 Too Many Requestswhen limit is exceeded
Logging
All login attempts are logged with timestamps and IP addresses:
✓ Successful login from 192.168.1.100- Successful authentication✗ Failed login attempt from 192.168.1.100- Invalid passwordLogin attempt without password from 192.168.1.100- Missing passwordLogin attempt when auth disabled from 192.168.1.100- Auth not configured
HTTP request logging (via morgan):
- Development: Concise colored output showing method, URL, status, and response time
- Production: Combined Apache-style logs with full details
Security Headers
Helmet middleware provides:
- Content Security Policy (CSP)
- X-Frame-Options (clickjacking protection)
- X-Content-Type-Options (MIME sniffing protection)
- Strict-Transport-Security (HTTPS enforcement in production)
- And other security headers
Best Practices
- Always set a strong
AUTH_PASSWORDfor deployments - Use a unique
SESSION_SECRETin production - Enable HTTPS in production (
NODE_ENV=production) - Monitor logs for suspicious login patterns
- Consider deploying behind a reverse proxy (nginx, Caddy) for additional security
♿ Accessibility Features
Codex is designed to be accessible to all users, including those using assistive technologies:
Screen Reader Support
- Comprehensive ARIA labels on all interactive elements
- Semantic HTML structure using
<header>,<nav>,<main>,<section>,<aside>, and<article>elements - Live regions (
aria-live) announce dynamic content updates - Proper roles (
role="tree",role="button",role="dialog") for enhanced navigation - Keyboard navigation support with proper focus management and
tabIndexattributes - Descriptive labels explain the state and purpose of all controls
Visual Accessibility
- Four theme options: Auto (follows system), Light, Dark, and High-Contrast
- High-contrast mode provides maximum visual clarity:
- Pure black (#000) background with white (#fff) text
- Yellow (#ffff00) secondary text for clear distinction
- Cyan (#00ffff) accent colors for links and interactive elements
- White borders for clear element separation
- Improved contrast ratios in all themes for WCAG compliance
- Larger interactive elements: Buttons sized at 32px for easier clicking
- Consistent theming: All features including reading mode support all themes
Keyboard Accessibility
- Tab navigation through all interactive elements
- Arrow key navigation (↑↓) or vim-style (
j/k) in folder tree, page list, and search results - Enter to activate buttons, open folders/pages, and select search results
- Escape to close modals and dialogs
- ⌘K/Ctrl+K global search shortcut
- Focus indicators show keyboard-selected items with blue outline
- Mouse hover sync updates keyboard selection for seamless interaction
The accessibility features ensure Codex can be used effectively by people with:
- Visual impairments (screen readers, high-contrast mode)
- Motor disabilities (keyboard-only navigation, larger click targets)
- Color blindness (semantic colors with sufficient contrast)
🚀 Quick Start
Installation
# Clone the repository
git clone https://github.com/bocan/codex.git
cd codex
# Install dependencies
make install
Configuration
Create a .env file in the root directory:
# Required: Set your password
AUTH_PASSWORD=your-secure-password-here
# Optional: Server port (default: 3001)
PORT=3001
Running
# Start both server and client
make dev
# Visit http://localhost:3000
# Login with your AUTH_PASSWORD
🛠️ Development
Project Scripts
# Development
make dev # Run both server and client
make dev-server # Run server only (port 3001)
make dev-client # Run client only (port 3000)
# Building
make build # Build both server and client
make build-server # Build server only
make build-client # Build client only
# Testing
make test # Run all tests
make test-server # Run backend tests
make test-client # Run frontend tests
# Maintenance
make clean # Remove all node_modules and build artifacts
make install # Fresh install of all dependencies
make help # Show all available commands
Development Workflow
- Make your changes in the appropriate files
- Tests run automatically in watch mode (if enabled)
- Server auto-reloads on file changes (via tsx)
- Client hot-reloads on file changes (via Vite HMR)
- Run
make testbefore committing
Git Workflow & Releases
This project uses Conventional Commits and commit-and-tag-version for automated changelog generation.
Commit Message Format
type(scope): description
Examples:
feat: add search functionality
fix: resolve dark mode flicker
docs: update README
refactor(api): simplify error handling
Types: feat, fix, docs, style, refactor, perf, test, build, ci, chore
Pull Request Workflow
# 1. Create and switch to feature branch
git switch -c feature/my-change
# 2. Make changes, commit (repeat as needed)
git add .
git commit -m "feat: add new feature"
# 3. Push branch to GitHub
git push -u origin feature/my-change
# 4. Create PR (via GitHub CLI or web)
gh pr create --title "feat: add new feature" --body "Description"
# 5. After PR is merged, switch back to main
git switch main
git pull
Creating a Release
After merging PRs and when ready to release:
# Patch release (1.0.0 → 1.0.1) - bug fixes
npm run release
# Minor release (1.0.0 → 1.1.0) - new features
npm run release:minor
# Major release (1.0.0 → 2.0.0) - breaking changes
npm run release:major
# Push with tags
git push --follow-tags
This will:
- Bump version in
package.json - Update
CHANGELOG.mdwith commits since last release - Create a git commit and tag (e.g.,
v1.1.0)
Adding New Features
Backend (Adding a new API endpoint)
- Create controller in
server/src/controllers/ - Add route in
server/src/routes/ - Register route in
server/src/index.ts - Add tests in
server/tests/
Frontend (Adding a new component)
- Create component in
client/src/components/ - Add corresponding CSS file
- Import and use in parent component
- Add tests in same directory (
.test.tsx)
🏗️ Project Structure
codex/
├── 📄 Makefile # Build automation
├── 📄 package.json # Root package config
├── 📄 README.md # This file
│
├── 📁 server/ # Backend (Express + TypeScript)
│ ├── 📄 package.json # Server dependencies
│ ├── 📄 tsconfig.json # TypeScript config
│ ├── 📄 jest.config.js # Test configuration
│ ├── 📁 src/
│ │ ├── 📄 index.ts # Server entry point
│ │ ├── 📁 controllers/ # Request handlers
│ │ │ ├── folderController.ts
│ │ │ └── pageController.ts
│ │ ├── 📁 routes/ # API routes
│ │ │ ├── folders.ts
│ │ │ └── pages.ts
│ │ └── 📁 services/ # Business logic
│ │ └── fileSystem.ts # File operations
│ └── 📁 tests/ # Test files
│ └── api.test.ts
│
├── 📁 client/ # Frontend (React + TypeScript)
│ ├── 📄 package.json # Client dependencies
│ ├── 📄 tsconfig.json # TypeScript config
│ ├── 📄 vite.config.ts # Vite configuration
│ ├── 📄 vitest.config.ts # Test configuration
│ ├── 📄 index.html # HTML entry point
│ ├── 📁 src/
│ │ ├── 📄 main.tsx # React entry point
│ │ ├── 📄 App.tsx # Main app component
│ │ ├── 📄 App.css # App styles
│ │ ├── 📁 components/ # React components
│ │ │ ├── FolderTree.tsx # Folder navigation
│ │ │ ├── FolderTree.css
│ │ │ ├── PageList.tsx # Page list in folder
│ │ │ ├── PageList.css
│ │ │ ├── Editor.tsx # Markdown editor
│ │ │ ├── Editor.css
│ │ │ ├── Preview.tsx # Markdown preview
│ │ │ ├── Preview.css
│ │ │ └── *.test.tsx # Component tests
│ │ ├── 📁 services/ # API client
│ │ │ └── api.ts
│ │ ├── 📁 types/ # TypeScript types
│ │ │ └── index.ts
│ │ └── 📁 test/ # Test setup
│ │ └── setup.ts
│
└── 📁 data/ # File storage
└── Welcome.md # Default welcome page
📋 Table of Contents
- Quick Start
- Installation
- Usage
- Project Structure
- AI Chat
- MCP Server (AI Agent Access)
- API Documentation
- Testing
- Development
- Tech Stack
- Makefile Commands
- Troubleshooting
- Contributing
🚀 Quick Start
The fastest way to get started using the Makefile:
# Install all dependencies
make install
# Run the application
make dev
Then open http://localhost:3000 in your browser!
📦 Installation
Prerequisites
- Node.js 25+ and npm
- macOS, Linux, or Windows with WSL
Option 1: Using Make (Recommended)
make install
Option 2: Using npm
npm run install:all
Option 3: Manual Installation
# Install root dependencies
npm install
# Install server dependencies
cd server && npm install
# Install client dependencies
cd ../client && npm install
🎯 Usage
Running the Application
Option 1: Using Make (Recommended)
# Run both server and client in development mode
make dev
# Or run them separately
make dev-server # Runs server on port 3001
make dev-client # Runs client on port 3000
Option 2: Using npm
# Run both concurrently
npm run dev
# Or run separately
npm run dev:server
npm run dev:client
The application will be available at:
- Frontend: http://localhost:3000
- Backend API: http://localhost:3001
- API Health Check: http://localhost:3001/api/health
Using the Web Interface
- Navigate Folders: Click on folders in the left pane to select them
- Create Folders: Right-click any folder and select "New Folder"
- Create Pages: Select a folder and click "+ New Page"
- Edit Content: Click on a page to open it in the editor (center pane)
- Rename Page: Right-click on a page and select "Rename"
- Move Page: Right-click on a page and select "Move to..." to move it to another folder
- Delete Page: Right-click on a page and select "Delete"
- Preview: See live markdown preview in the right pane
- Auto-save: Content saves automatically after 2 seconds
- Resize Panes:
- Drag the edge of left/right panes to resize horizontally
- Drag the divider between folder tree and page list to resize vertically
- Collapse Panes: Use the arrow buttons to hide left/right panes
- Theme Toggle: Click the theme button (🌓/☀️/🌙/◐) in the header to cycle through auto/light/dark/high-contrast modes
Production Build
# Using Make
make build
# Using npm
npm run build
This creates optimized production builds in:
server/dist/- Compiled backendclient/dist/- Optimized frontend bundle
� Docker Deployment
Codex can be run in Docker for simplified deployment. In production mode, Express serves both the API and the React UI from a single port (3001).
Pre-built Images
Official multi-architecture (AMD64/ARM64) images are available on Docker Hub:
docker pull bocan/codex:latest
Available tags:
latest- Latest stable release2.16.0- Specific version (example)2.16- Minor version (example)2- Major version (example)
Image verification (signed with Cosign):
cosign verify bocan/codex:latest \
--certificate-identity-regexp=https://github.com/bocan/codex \
--certificate-oidc-issuer=https://token.actions.githubusercontent.com
Supply chain security:
- Images are signed with Cosign for authenticity verification
- SBOM (Software Bill of Materials) embedded in image metadata
- Provenance attestations document the build process
To inspect the SBOM:
docker buildx imagetools inspect bocan/codex:latest --format "{{ json .SBOM }}"
Using Docker Compose (Recommended)
# Build and start the container
docker compose up -d --build
# View logs
docker compose logs -f codex
# Stop the container
docker compose down
The application will be available at http://localhost:3001 (both UI and API on the same port).
Using Raw Docker
# Build the image
docker build -t codex .
# Run the container
docker run -d \
--name codex \
-p 3001:3001 \
-v $(pwd)/data:/app/data \
-e NODE_ENV=production \
codex
# View logs
docker logs -f codex
# Stop and remove
docker stop codex && docker rm codex
Docker Configuration
Environment Variables (in docker-compose.yml):
NODE_ENV=production- Runs in production modePORT=3001- Server port (default)AUTH_PASSWORD=your-password- Set to enable authentication (optional)TRUST_PROXY=true- Enable when behind a reverse proxy with HTTPS
Volume Mounting:
./data:/app/data- Persists your wiki data outside the container
Templates on first run:
- The Docker image includes built-in starter templates; on first startup, if
data/templates/is missing or empty, Codex will copy the defaults into it.
Authentication in Docker
Important: Due to secure cookie requirements, there are two ways to run Codex with authentication:
Option 1: Run Passwordless Locally (for testing)
Remove or comment out the AUTH_PASSWORD environment variable in docker-compose.yml:
environment:
- NODE_ENV=production
- PORT=3001
# - AUTH_PASSWORD=your-password # Commented out
This allows direct HTTP access without authentication, suitable for local testing.
Option 2: Deploy Behind a Reverse Proxy (for production)
For production with authentication, deploy behind a reverse proxy (nginx, Caddy, Traefik) that:
- Terminates TLS/HTTPS
- Forwards requests to Codex with the
X-Forwarded-Proto: httpsheader - Set
TRUST_PROXY=truein the environment variables
Example with nginx:
server {
listen 443 ssl;
server_name wiki.example.com;
ssl_certificate /path/to/cert.pem;
ssl_certificate_key /path/to/key.pem;
location / {
proxy_pass http://localhost:3001;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
}
}
Why? In production mode, session cookies use secure: auto, which requires HTTPS. Direct HTTP access with a password set will fail because the browser won't send the secure cookie. The reverse proxy provides HTTPS termination while communicating with Codex via HTTP internally.
🤖 MCP Server (AI Agent Access)
Codex includes a Model Context Protocol (MCP) server that allows AI agents like Claude, GitHub Copilot, and other MCP-compatible clients to interact with your documentation.
Quick Setup
-
Enable the MCP server by setting environment variables:
export MCP_ENABLED=true export MCP_API_KEY=your-secure-api-key -
Start the server (runs alongside the main app):
npm run dev:mcp -w server # Development with hot reload -
Connect your MCP client to
http://localhost:3002/mcp
Available Tools
The MCP server exposes 16 tools for AI agents:
| Tool | Description |
|---|---|
search_pages |
Search documentation by query |
get_page |
Read a page's content |
create_page |
Create a new page |
update_page |
Update an existing page |
delete_page |
Delete a page |
rename_page |
Rename a page |
move_page |
Move a page to another folder |
list_folders |
Get folder hierarchy |
list_pages |
List pages in a folder |
create_folder |
Create a new folder |
delete_folder |
Delete an empty folder |
rename_folder |
Rename a folder |
list_attachments |
List attachments in a folder |
upload_attachment |
Upload a file attachment |
get_attachment |
Download an attachment |
delete_attachment |
Delete an attachment |
Questions sur Codex MCP
Quels outils Codex MCP fournit-il ?
Il expose 16 outils, dont search_pages, get_page, create_page, update_page, delete_page.
Puis-je utiliser des modèles IA locaux avec Codex ?
Oui, vous pouvez utiliser Ollama pour exécuter des modèles locaux comme Llama ou Mistral en configurant l'hôte et le port dans les paramètres.
Comment Codex gère-t-il le versionnage des documents ?
Il utilise un contrôle de version basé sur Git, ce qui vous permet de visualiser les diffs et de restaurer des versions précédentes de vos documents.