Connects Claude to PullPush.io's historical Reddit data through two search operations: one for comments and one for submissions. You get comprehensive filtering by subreddit, author, date ranges, score thresholds, and post attributes like NSFW status or video content. Useful when you need to analyze Reddit discussions beyond what the official API exposes, search deleted or archived content, or pull historical data without OAuth setup. Supports both epoch timestamps and relative date formats like "30d" or "1y". Returns up to 100 results per query with sorting by timestamp, score, or comment count.
An MCP (Model Context Protocol) server for the PullPush.io Reddit API. This allows Claude and other MCP-compatible AI assistants to search Reddit comments and submissions.
PullPush.io provides access to Reddit's historical data, making it possible to search posts and comments that may no longer be available through Reddit's official API.
search_commentsSearch Reddit comments with the following parameters:
| Parameter | Type | Description |
|---|---|---|
q | string | Search query across all comment fields |
subreddit | string | Filter by subreddit (without r/ prefix) |
author | string | Filter by username |
after | string | Results after date (epoch or relative: 30d, 1y) |
before | string | Results before date (epoch or relative: 30d, 1y) |
sort | asc | desc | Sort order (default: desc) |
sort_type | created_utc | score | Sort field (default: created_utc) |
size | number | Results to return (1-100, default: 100) |
search_submissionsSearch Reddit posts/submissions with the following parameters:
| Parameter | Type | Description |
|---|---|---|
q | string | Search query across all fields |
subreddit | string | Filter by subreddit (without r/ prefix) |
author | string | Filter by username |
title | string | Search in titles only |
selftext | string | Search in post body only |
after | string | Results after date (epoch or relative: 30d, 1y) |
before | string | Results before date (epoch or relative: 30d, 1y) |
sort | asc | desc | Sort order (default: desc) |
sort_type | created_utc | score | num_comments | Sort field |
size | number | Results to return (1-100, default: 25) |
score | string | Filter by score (>100, <50, or exact) |
num_comments | string | Filter by comment count (>10, <5, or exact) |
over_18 | boolean | Filter NSFW content |
is_video | boolean | Filter video posts |
locked | boolean | Filter locked posts |
stickied | boolean | Filter stickied posts |
spoiler | boolean | Filter spoiler posts |
npm install -g pullpush-mcp
git clone https://github.com/jacklenzotti/pullpush-mcp.git
cd pullpush-mcp
npm install
npm run build
Add to your Claude Desktop config file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Using npm (recommended):
{
"mcpServers": {
"pullpush": {
"command": "npx",
"args": ["-y", "pullpush-mcp"]
}
}
}
Using a local build:
{
"mcpServers": {
"pullpush": {
"command": "node",
"args": ["/absolute/path/to/pullpush-mcp/build/index.js"]
}
}
}
Restart Claude Desktop after updating the config.
Once configured, you can ask Claude:
# Build
npm run build
# Type check
npm run typecheck
# Test with MCP Inspector
npm run inspect
This server uses the PullPush.io API. PullPush provides free access to Reddit's historical data without requiring authentication.
MIT
com.mcparmory/google-search
io.github.pipeworx-io/brave-search
marcopesani/mcp-server-serper
brave/brave-search-mcp-server
com.mcparmory/google-search-console
acamolese/google-search-console-mcp