reddit-mcp-server
🔗 A zero-config Model Context Protocol (MCP) server for Reddit — search posts, browse subreddits, scrape user activity, and get structured data via any MCP-compatible AI client. No API keys or authentication needed.
Links
README
From the repo.
Reddit MCP Server
A Model Context Protocol (MCP) server that provides AI assistants with access to Reddit data. Built on top of redd — no API keys required.
https://github.com/user-attachments/assets/af8ae52c-f9f3-4d04-80d5-4be0dfa61b0f
Table of Contents
1. Features
- 🔍 Search — Search all of Reddit or within a specific subreddit
- 📰 Subreddit Posts — Browse hot, top, new, or rising posts from any subreddit
- 📖 Post Details — Get full post content with nested comment trees
- 👤 User Activity — View a user's recent posts and comments
- 📝 User Posts — Get a user's submitted posts
No API keys, no authentication, no browser required. Just install and run.
2. Quick Start
2.1. Using uvx (recommended)
The fastest way to run the server — no clone needed:
# stdio transport (default, for Claude Desktop / Cursor / etc.)
uvx reddit-no-auth-mcp-server
# HTTP transport
uvx reddit-no-auth-mcp-server \
--transport streamable-http \
--port 8000
2.2. From source
git clone https://github.com/eliasbiondo/reddit-mcp-server.git
cd reddit-mcp-server
uv sync
Run the server:
# stdio transport (default)
uv run reddit-no-auth-mcp-server
# HTTP transport
uv run reddit-no-auth-mcp-server \
--transport streamable-http \
--port 8000
2.3. MCP Client Configuration
Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"reddit": {
"command": "uvx",
"args": ["reddit-no-auth-mcp-server"]
}
}
}
Cursor
Add to your .cursor/mcp.json:
{
"mcpServers": {
"reddit": {
"command": "uvx",
"args": ["reddit-no-auth-mcp-server"]
}
}
}
From source (any MCP client)
{
"mcpServers": {
"reddit": {
"command": "uv",
"args": [
"--directory", "/path/to/reddit-mcp-server",
"run", "reddit-no-auth-mcp-server"
]
}
}
}
3. Available Tools
| Tool | Description | Key Arguments |
|---|---|---|
search | Search Reddit for posts | query, limit, sort |
search_subreddit | Search within a subreddit | subreddit, query, limit, sort |
get_post | Get post details + comment tree | permalink |
get_subreddit_posts | Get subreddit listing | subreddit, limit, category, time_filter |
get_user | Get user's activity feed | username, limit |
get_user_posts | Get user's submitted posts | username, limit, category, time_filter |
Tool Details
search
Search all of Reddit for posts matching a query.
query: "python async programming"
limit: 10
sort: "relevance" # relevance, hot, top, new, comments
search_subreddit
Search within a specific subreddit.
subreddit: "Python"
query: "web scraping"
limit: 10
sort: "top"
get_post
Get full details of a Reddit post including its comment tree.
permalink: "/r/Python/comments/abc123/my_post/"
get_subreddit_posts
Get posts from a subreddit listing.
subreddit: "MachineLearning"
limit: 25
category: "hot" # hot, top, new, rising
time_filter: "week" # hour, day, week, month, year, all
get_user
Get a user's recent public activity (posts and comments).
username: "spez"
limit: 10
get_user_posts
Get a user's submitted posts.
username: "spez"
limit: 10
category: "top" # hot, top, new
time_filter: "all" # hour, day, week, month, year, all
4. Configuration
All settings can be configured via environment variables:
| Variable | Default | Description |
|---|---|---|
REDDIT_TRANSPORT | stdio | MCP transport (stdio, streamable-http) |
REDDIT_HOST | 127.0.0.1 | Host for HTTP transport |
REDDIT_PORT | 8000 | Port for HTTP transport |
REDDIT_PATH | /mcp | Path for HTTP transport |
REDDIT_LOG_LEVEL | WARNING | Log level (DEBUG, INFO, WARNING, ERROR) |
REDDIT_PROXY | — | HTTP/HTTPS proxy URL |
REDDIT_TIMEOUT | 10.0 | Request timeout in seconds |
REDDIT_THROTTLE_MIN | 1.0 | Min delay between paginated requests (seconds) |
REDDIT_THROTTLE_MAX | 2.0 | Max delay between paginated requests (seconds) |
CLI arguments take precedence over environment variables:
uv run reddit-no-auth-mcp-server \
--transport streamable-http \
--port 9000 \
--log-level DEBUG
5. Architecture
This project follows hexagonal architecture (ports & adapters):
src/reddit_mcp_server/
├── domain/ # Pure business logic, no framework imports
│ ├── exceptions.py # Domain exception hierarchy
│ └── value_objects.py # Immutable config objects
├── ports/ # Abstract interfaces (contracts)
│ ├── config.py # ConfigPort
│ └── reddit.py # RedditPort
├── application/ # Use cases (orchestration)
│ ├── search.py
│ ├── search_subreddit.py
│ ├── get_post.py
│ ├── get_user.py
│ ├── get_subreddit_posts.py
│ └── get_user_posts.py
├── adapters/
│ ├── inbound/ # Presentation layer
│ │ ├── cli.py # CLI entry point
│ │ ├── mcp_server.py
│ │ ├── error_mapping.py
│ │ ├── serialization.py
│ │ └── mcp_tools/ # MCP tool definitions
│ └── outbound/ # Infrastructure layer
│ ├── env_config.py # ConfigPort implementation
│ └── redd_client.py # RedditPort implementation (wraps redd)
└── container.py # DI composition root
6. Contributing
Contributions are welcome. Please read CONTRIBUTING.md for guidelines on setting up the project, running tests, and submitting changes.
7. License
Collected info
- ★ 144 stars
- ⎇ 15 forks
- Language: Python
- Source updated: 7/29/2026
Config for your environment
Replace {MCP_ENDPOINT_URL} with this MCP’s endpoint URL (from its repo or docs above). No API key — you connect directly.
Tool
OS
Config file: ~/.cursor/mcp.json
{
"mcpServers": {
"mcp-server": {
"url": "{MCP_ENDPOINT_URL}"
}
}
}Paste into mcpServers in the config file. Restart Cursor after saving.
If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.