Complete reference for Convext's Model Context Protocol (MCP) server and REST API.
Authentication
All API requests require authentication via Bearer token.
Creating an API Token
- Log in to Convext web UI
- Navigate to Settings → API Tokens
- Click New Token
- Enter a description (e.g., "CLI Access", "Claude Code MCP")
- Click Create
- Copy the token immediately (shown only once)
Using the Token
MCP Overview
The Model Context Protocol (MCP) server provides structured access to Convext data for AI coding assistants.
MCP Server Endpoints
| Endpoint | Method | Description |
|---|---|---|
| /mcp | GET | Server manifest (capabilities discovery) |
| /mcp | POST | JSON-RPC method calls |
MCP Tools
Tools are callable functions the AI can invoke.
get_rules
Get all engineering rules and guidelines.
Input: None required
Use Cases: "What are all the engineering rules?", "Show me critical security rules"
get_tech_standards
Get technology standards (language, framework, tooling setup).
Input: Optional languages filter (comma-separated)
Use Cases: "What's the Python testing standard?", "Show me all Ruby standards"
register_project
Register a new project with Convext.
Use Cases: "Register this repo with Convext"
get_project_context
Get full context for the current project including rules, standards, and metadata.
Use Cases: "What are the rules for this project?"
MCP Resources
Resources are data the AI can read directly.
| Resource URI | Description |
|---|---|
| charter:///rules | All engineering rules grouped by category |
| charter:///rulesets | Available rulesets with metadata |
| charter:///standards | All language standards with tech details |
REST API
Direct HTTP access to Convext data.
Get Global Rules
Get Tech Standards
List Projects
Get Project Context
Error Handling
401 Unauthorized
Invalid or missing API token. Check your Authorization header.
404 Not Found
Resource does not exist. Verify project ID or resource URI.
429 Too Many Requests
Rate limit exceeded. Wait before retrying.