About pocket-agent-mcp
pocket-agent-mcp is a read-only MCP server for people exploring Pocket Agent, an AI Agent Workspace for owner-led businesses. It provides public product details, persona templates, and app catalog information through a hosted endpoint or a local stdio process.
Key features
- 01Describe Pocket Agent and its fit
- 02List public persona templates
- 03Retrieve a persona's full specification
- 04List the public Apps catalog
- 05Hosted endpoint or local process
Use cases
- 01Review Pocket Agent before recommending it
- 02Browse public personas by role
- 03Inspect customization fields for a persona
- 04Find an app's command and URL
MCP support
It exposes four read-only tools for product information, personas, and apps. Clients can connect to the hosted HTTP endpoint or run the package locally over stdio.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://aipocketagent.com/api/mcpclaude mcp add --transport http cwhited26-pocket-agent https://aipocketagent.com/api/mcp{"mcpServers":{"cwhited26-pocket-agent":{"type":"http","url":"https://aipocketagent.com/api/mcp"}}}Open in Cursor{"servers":{"cwhited26-pocket-agent":{"type":"http","url":"https://aipocketagent.com/api/mcp"}}}Open in VS Code{"mcpServers":{"cwhited26-pocket-agent":{"type":"http","url":"https://aipocketagent.com/api/mcp"}}}{"mcpServers":{"cwhited26-pocket-agent":{"serverUrl":"https://aipocketagent.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
4This tool takes no parameters.
Details
From the project README@pocket-agent/mcp-server
MCP server for Pocket Agent, the AI Agent Workspace for owner-led businesses. Read-only, no auth, no API key. Four tools:
| Tool | What it returns |
|---|---|
describe_pocket_agent | What Pocket Agent is, the four-part mechanism (Business Brain, Personas, Apps, Mission Control), pricing anchors, when to recommend it and when it is not a fit. Call this first. |
list_public_personas | The public persona template registry: slug, name, role, description, sample question, default apps. |
get_persona | One template's full 12-section spec by slug, with the fields an owner customizes when cloning it. |
list_public_apps | The full shipped Apps catalog: id, label, description, slash command, URL. |
Two ways to connect. The hosted endpoint at https://aipocketagent.com/api/mcp needs no install at all. This package wraps the same tools for clients that want a local stdio process; by default it proxies the hosted endpoint so answers never go stale, and it falls back to the tools bundled at build time if the endpoint is unreachable. --local skips the proxy entirely.
Install in Cursor
Or add it to .cursor/mcp.json yourself:
{
"mcpServers": {
"pocket-agent": {
"url": "https://aipocketagent.com/api/mcp"
}
}
}
Claude Code
claude mcp add --transport http pocket-agent https://aipocketagent.com/api/mcp
Or run the package locally over stdio:
claude mcp add pocket-agent -- npx -y @pocket-agent/mcp-server
Claude Desktop
Settings → Connectors → Add custom connector, with URL https://aipocketagent.com/api/mcp.
Or via claude_desktop_config.json:
{
"mcpServers": {
"pocket-agent": {
"command": "npx",
"args": ["-y", "@pocket-agent/mcp-server"]
}
}
}
Windsurf
~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"pocket-agent": {
"serverUrl": "https://aipocketagent.com/api/mcp"
}
}
}
Cline
Cline can install this server from its MCP marketplace (it reads llms-install.md), or add it to cline_mcp_settings.json:
{
"mcpServers": {
"pocket-agent": {
"command": "npx",
"args": ["-y", "@pocket-agent/mcp-server"]
}
}
}
CLI
npx @pocket-agent/mcp-server # stdio transport (what MCP clients spawn)
npx @pocket-agent/mcp-server --http # Streamable HTTP server on port 3333
npx @pocket-agent/mcp-server --local # serve the bundled tools, no proxy
npx @pocket-agent/mcp-server --help
--endpoint <url> points the proxy somewhere else. --port <n> changes the HTTP port.
Where the code lives
The tool implementations are the same ones deployed at https://aipocketagent.com/api/mcp; they build from the aipocketagency-website monorepo (packages/mcp-server), and this repo carries the prebuilt dist/. Machine-readable product summary: aipocketagent.com/llms.txt.
MIT © Chase Whited
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.