dynamic-discord-bot-api-mcp
MCP server for Discord bot API that exposes five tools to search, inspect, and call stable Discord HTTP endpoints using bot-token authentication, with automatic schema refresh and strict safety filtering.
README
dynamic-discord-bot-api-mcp
Production-oriented stdio MCP gateway exposing stable Discord HTTP operations that explicitly support bot-token authentication. It installs a checked-in, self-refreshing OpenAPI catalog and exposes exactly five generic tools—never one tool per endpoint.
Discord's OpenAPI specification is a public preview and may be incomplete. This project consumes only stable openapi.json; official Discord documentation remains authoritative.
Scope and architecture
Refresh filters every operation by explicit BotToken security metadata, then applies a small fail-closed classifier for OAuth, interaction-token, webhook-token, Social SDK/lobby, and ambiguous routes. It retains only transitively referenced component schemas. Excluded endpoints cannot be searched, inspected, or called.
Gateway HTTP metadata such as /gateway/bot may remain because it supports bot authentication; Gateway WebSockets and events do not. Also excluded: OAuth exchange, bearer/user authorization, self-bots, Voice WebSocket/RTP, RPC, Activities/Embedded App/Social SDKs, interaction receivers/callbacks, incoming webhook servers, webhook-token execution, preview endpoints, and arbitrary HTTP requests.
Runtime resolves one active local MCP workspace root, reloads its .env before every call, and caches a client only while security-relevant configuration is unchanged. Schema refresh uses validation, atomic rename, concurrent coalescing, and cached fallback. JSON audit events go to stderr without values or secrets.
Install
npm install -g dynamic-discord-bot-api-mcp
discord-bot-api-mcp
From GitHub:
npm install -g github:PrimeUpYourLife/dynamic-discord-bot-api-mcp
MCP client configuration:
{"mcpServers":{"discord":{"command":"discord-bot-api-mcp"}}}
Codex ~/.codex/config.toml:
[mcp_servers.discord]
command = "discord-bot-api-mcp"
Place DISCORD_BOT_TOKEN in active repository .env. Process environment overrides startup-directory .env; active workspace .env overrides both. Multiple, failed, non-local, or non-directory MCP roots fail closed. Catalog tools need no token.
Five tools
discord_search_endpoints: fuzzy search IDs, methods, routes, categories, descriptions, and parameters.discord_get_endpoint: exact canonical ID or OpenAPI operation-ID lookup.discord_get_type: component-schema lookup.discord_call_endpoint: validate and execute a cataloged bot endpoint.discord_refresh_schema: fetch, filter, validate, and atomically install stable schema.
Examples:
{"search":"create message","limit":5}
{"endpoint":"create_message"}
{"endpoint":"POST /channels/{channel_id}/messages","pathParameters":{"channel_id":"123456789012345678"},"body":{"content":"Hello from MCP"}}
Upload example:
{"endpoint":"create_message","pathParameters":{"channel_id":"123456789012345678"},"body":{"content":"image"},"files":[{"path":"./uploads/image.png","filename":"image.png","contentType":"image/png"}]}
Uploads use payload_json and deterministic files[n] parts/attachment IDs. Paths resolve from active root and must remain within configured realpath roots. Base64 descriptors are also supported.
Validation separates path, query, and body values; handles required fields, nested objects, arrays, enums, nullable values, refs and unions. Snowflakes remain strings. Arrays become repeated query keys. DELETE and centrally classified destructive operations require confirm: true. Optional audit reason is percent-encoded.
Errors are structured (ENDPOINT_NOT_FOUND, ENDPOINT_NOT_ALLOWED, CONFIRMATION_REQUIRED, BOT_TOKEN_MISSING, VALIDATION_ERROR, DISCORD_API_ERROR, REQUEST_ERROR, project-root errors, or SCHEMA_REFRESH_FAILED). Discord numeric codes and nested validation errors are preserved; raw bodies, stack traces, and tokens are not.
Configuration
| Variable | Default | Meaning |
|---|---|---|
DISCORD_BOT_TOKEN |
unset | Bot token; calls fail safely when absent |
DISCORD_API_BASE_URL |
https://discord.com/api/v10 |
Fixed API origin |
DISCORD_ENDPOINT_ALLOWLIST |
* |
Comma-separated IDs/operation IDs with * globs |
DISCORD_REQUEST_TIMEOUT_MS |
30000 |
Per-attempt timeout |
DISCORD_REQUEST_RETRIES |
2 |
Retry count |
DISCORD_RATE_LIMIT_PER_SECOND |
45 |
Local global safety rate |
DISCORD_RATE_LIMIT_BURST |
50 |
Local burst |
DISCORD_SCHEMA_MAX_AGE_HOURS |
24 |
Refresh threshold |
DISCORD_SCHEMA_PATH |
bundled catalog | Alternate catalog path |
DISCORD_LOCAL_FILE_ROOTS |
active root | Platform-delimited upload roots |
DISCORD_MAX_UPLOAD_BYTES |
10485760 |
Per-file limit |
DISCORD_ALLOW_UNKNOWN_PARAMETERS |
false |
Validation escape hatch |
LOG_LEVEL |
info |
debug, info, warn, error |
Use narrow production allowlists. Caller cannot supply URLs, methods, authorization, or headers. Requests always use configured base URL plus catalog route, Authorization: Bot …, and valid package User-Agent.
Discord response headers and 429 payloads control per-route/bucket and global waits; limits are not hard-coded. Fractional retry_after is honored. 429s retry within configured bound. 5xx and transport retries are limited to safe GET requests, preventing duplicate messages or moderation actions.
Development
npm ci
npm run refresh-schema
npm run check
npm test
npm run build
Tests mock network access. Daily workflow refreshes catalog and commits only generated-data changes. Manual publish workflow verifies tag/version and uses npm trusted publishing via GitHub OIDC; prereleases publish under next.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.