About Tweetsmash
Tweetsmash provides an MCP server for searching and managing X bookmarks in a saved-post library. It supports keyword and semantic search, bookmark counts, and label operations, with a hosted endpoint available as well as a local package.
Key features
- 01Keyword and semantic bookmark search
- 02Browse bookmarks with filters
- 03Count bookmarks by label
- 04List and manage labels
- 05Build a semantic search index
Use cases
- 01Find saved posts about a topic
- 02Count bookmarks associated with a label
- 03Apply labels to a group of saved posts
MCP support
The server exposes bookmark search and management tools to MCP clients. It can run through the @tweetsmash/mcp package or the hosted endpoint at https://mcp.tweetsmash.com/api/mcp.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.tweetsmash.com/api/mcpclaude mcp add --transport http tweetsmash https://mcp.tweetsmash.com/api/mcp{"mcpServers":{"tweetsmash":{"type":"http","url":"https://mcp.tweetsmash.com/api/mcp"}}}Open in Cursor{"servers":{"tweetsmash":{"type":"http","url":"https://mcp.tweetsmash.com/api/mcp"}}}Open in VS Code{"mcpServers":{"tweetsmash":{"type":"http","url":"https://mcp.tweetsmash.com/api/mcp"}}}{"mcpServers":{"tweetsmash":{"serverUrl":"https://mcp.tweetsmash.com/api/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
5| Parameter | Type | Required | Description |
|---|---|---|---|
| limit | number | No | Number of bookmarks to retrieve (1-100) |
| cursor | string | No | Pagination cursor for next page of results |
| is_unread_only | boolean | No | Filter to only unread bookmarks |
| hide_archived | boolean | No | Hide archived bookmarks from results |
| media_type | enum | No | Filter by specific media typethread | media | replies | notes | links |
| author | string | No | Filter by specific author username |
| tag | string | No | Filter by specific tag |
| posted_from | string | No | Filter tweets posted after this date (ISO 8601 format) |
| posted_to | string | No | Filter tweets posted before this date (ISO 8601 format) |
| bookmarked_from | string | No | Filter bookmarks added after this date (ISO 8601 format) |
| bookmarked_to | string | No | Filter bookmarks added before this date (ISO 8601 format) |
| sort_by | enum | No | Sort order for resultsPOSTED_DATE_ASC | POSTED_DATE_DESC | OLDEST_FIRST | NEWEST_FIRST |
Details
From the project READMETweetsmash MCP Server: X (Twitter) Bookmarks for AI Agents
The Tweetsmash MCP server connects Claude, ChatGPT, Cursor, Raycast, and any Model Context Protocol client to your X (Twitter) bookmarks. Search them semantically, count them, and manage labels through natural language.
Tweetsmash is a bookmark manager for X (Twitter): it imports the posts you bookmark on X into a searchable library, syncs them to Notion, Google Sheets, and Zotero, and exposes them to AI agents through this MCP server and an X bookmarks API.
Quick start
- Create a Tweetsmash account at tweetsmash.com and import your X bookmarks (one click).
- Generate an API token: Settings > Integrations > API.
- Add the server to your MCP config:
{
"mcpServers": {
"tweetsmash": {
"command": "npx",
"args": ["@tweetsmash/mcp"],
"env": { "TWEETSMASH_TOKEN": "your-token-here" }
}
}
}
Hosted endpoint (no local install)
Point any remote MCP client at:
https://mcp.tweetsmash.com/api/mcp
with your Tweetsmash API token as a Bearer header.
Tools
| Tool | What it does |
|---|---|
list_bookmarks | Page through saved posts with filters |
search_bookmarks | Keyword and semantic search across every bookmark |
get_bookmark_count | Count bookmarks, optionally by label |
list_labels | List every label in your library |
add_labels_to_tweets | Label bookmarks from your agent |
remove_labels_from_tweets | Remove labels |
setup_vector_store | Build the semantic search index |
Example prompts
- "Find my bookmarks about prompt engineering"
- "How many bookmarks mention MCP?"
- "Label everything I saved last week as AI"
Links
- X Bookmarks MCP
- Setup guide: Claude Desktop, Raycast, Smithery
- X Bookmarks API
- npm: @tweetsmash/mcp
- Smithery listing
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.