About AgentMash
AgentMash gathers bookmarks synced from a browser extension into one library that AI agents can search read-only. It is for people who want agents to consult saved items, sources, and collections.
Key features
- 01Collect bookmarks from multiple platforms
- 02Search saved items
- 03Retrieve full details for a save
- 04List sources and collections
- 05Read-only access
Use cases
- 01Search previously saved social posts
- 02View a save’s text, links, and media
- 03Check saved items and sync time by platform
- 04Retrieve bookmarks from a collection
MCP support
The hosted MCP server exposes six read-only tools over Streamable HTTP and authenticates with an API key. Connected clients can search and read the AgentMash library.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://agentmash.ai/mcpclaude mcp add --transport http ai-agentmash-agentmash https://agentmash.ai/mcp{"mcpServers":{"ai-agentmash-agentmash":{"type":"http","url":"https://agentmash.ai/mcp"}}}Open in Cursor{"servers":{"ai-agentmash-agentmash":{"type":"http","url":"https://agentmash.ai/mcp"}}}Open in VS Code{"mcpServers":{"ai-agentmash-agentmash":{"type":"http","url":"https://agentmash.ai/mcp"}}}{"mcpServers":{"ai-agentmash-agentmash":{"serverUrl":"https://agentmash.ai/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.
https://agentmash.ai/mcpAuth requiredTools
Details
From the project READMEAgentMash MCP server
All your bookmarks in one library. Share them with your AI agents.
This repo documents the hosted AgentMash MCP server: how to connect your AI agent to your AgentMash library.
What AgentMash does
- The AgentMash Chrome extension imports your saves from X, LinkedIn, Instagram and more when you click Sync. Nothing runs in the background.
- Everything lands in one library.
- Your AI agents read that library over MCP, with read-only tools.
Server
- URL:
https://agentmash.ai/mcp - Transport: Streamable HTTP
Authentication
AgentMash uses an API key, not OAuth.
- In AgentMash, open the Agents & API page and click Create API key.
- Copy the key right away. It starts with
am_live_and is shown only once. - Send it as a header:
Authorization: Bearer am_live_YOUR_KEY
For clients that cannot set headers, append ?token=am_live_YOUR_KEY to the URL instead.
Setup
Replace am_live_YOUR_KEY with your own key in each snippet.
Claude Code
claude mcp add --transport http agentmash https://agentmash.ai/mcp --header "Authorization: Bearer am_live_YOUR_KEY"
Claude Desktop
Claude Desktop connects through the mcp-remote bridge, so you need Node (with npx) installed. Add this to claude_desktop_config.json:
{
"mcpServers": {
"agentmash": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://agentmash.ai/mcp",
"--header",
"Authorization: Bearer am_live_YOUR_KEY"
]
}
}
}
Cursor
Add this to ~/.cursor/mcp.json:
{
"mcpServers": {
"agentmash": {
"url": "https://agentmash.ai/mcp",
"headers": {
"Authorization": "Bearer am_live_YOUR_KEY"
}
}
}
}
Cline
In Cline, open MCP Servers, then Configure, then Configure MCP Servers, and add this to cline_mcp_settings.json. Set "type": "streamableHttp": without it Cline falls back to the legacy SSE transport and will not connect.
{
"mcpServers": {
"agentmash": {
"type": "streamableHttp",
"url": "https://agentmash.ai/mcp",
"headers": {
"Authorization": "Bearer am_live_YOUR_KEY"
},
"disabled": false,
"autoApprove": []
}
}
}
Or with the Cline CLI:
cline mcp add agentmash https://agentmash.ai/mcp --transport http --header "Authorization: Bearer am_live_YOUR_KEY" --yes
Codex
Add this to ~/.codex/config.toml:
[mcp_servers.agentmash]
url = "https://agentmash.ai/mcp"
bearer_token_env_var = "AGENTMASH_API_KEY"
Then set the key in your shell before starting Codex: export AGENTMASH_API_KEY=am_live_YOUR_KEY. Codex reads the bearer token from that environment variable.
Tools
All six tools are read-only.
| Tool | What it does |
|---|---|
search_saves | Searches your saves. Leave the query out to list your newest saves first. |
get_save | Gets one save in full: text, stats, links, media and collections. |
recent_saves | Lists saves from the last few days. |
list_sources | Lists the platforms in your library, with saves per platform and last sync. |
list_collections | Lists your collections with their names, platform and number of saves. |
get_collection_saves | Gets the saves inside one collection. |
Revoke a key
Open the Agents & API page in AgentMash and click Revoke on the key. Agents using that key stop working within about a minute.
Links
- AgentMash
- MCP setup guide: https://agentmash.ai/mcp-setup
About this repo
The AgentMash app source code is not in this repo. This repo only documents the hosted MCP server.
For AI coding agents, see llms-install.md.
Support
Email [email protected].
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.