mcpField
L
Lycatra

Lycatra

Remote
lycatra.comProductivity0Updated 5d ago

About Lycatra

Lycatra connects MCP-enabled agent apps to account features such as teams, schedules, notifications, memory, wiki, Matrix chat, and vault metadata. Individuals and teams can access the same account through a hosted MCP server or local CLI, with visibility limited by account permissions.

Key features

  1. 01Access teams, schedules, and notifications
  2. 02Use memory and wiki
  3. 03Connect Matrix chat and vault metadata
  4. 04Sign in with OAuth
  5. 05Use the local Lycatra CLI

Use cases

  1. 01Access team information from agent apps
  2. 02Manage account schedules and notifications
  3. 03Run computer-side tasks through the local CLI

MCP support

Lycatra’s hosted MCP server at `https://lycatra.com/api/mcp` exposes account features to clients and uses OAuth for authorization. The Lycatra CLI supports the same account operations in authorized local or cloud shells.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://lycatra.com/api/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

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

Details

From the project README

Lycatra plugin

Lycatra for Claude, ChatGPT, Codex, Cursor, VS Code, and any other app that speaks MCP: your agents, wiki, memory, teams, schedules, notifications, Matrix chat, and vault, in one sign-in.

Install

Every app below connects to https://lycatra.com/api/mcp and asks you to sign in to Lycatra once. One-click buttons for every app are on lycatra.com/connect.

Claude

Claude on the web, desktop, and mobile, and Cowork.

Recommended: the plugin, with skills:

  1. In Claude, open Customize, then Plugins, then Add marketplace.
  2. Enter Lycatra/lycatra-plugin and turn on Sync automatically.
  3. Install Lycatra, then connect it when Claude asks you to sign in.

A plugin added here also appears in Claude Code when you sign in with the same account.

Team and Enterprise plans: an organization owner adds Lycatra, then members connect it.

ChatGPT

ChatGPT on the web and desktop.

Developer mode:

  1. In ChatGPT, open Settings, then Security and login, and turn on Developer mode.
  2. Open chatgpt.com/plugins, choose the plus button, and enter https://lycatra.com/api/mcp.
  3. Sign in to Lycatra when ChatGPT asks.

Until Lycatra is in the ChatGPT directory, adding it needs developer mode, which depends on your plan.

Claude Code

Lycatra tools, plus the Lycatra CLI for work that needs your computer.

claude plugin marketplace add Lycatra/lycatra-plugin
claude plugin install lycatra@lycatra

Turn on updates in /plugin, Marketplaces, then Enable auto-update. lycatra setup does this for you.

Codex

The Codex CLI, the Codex app, and ChatGPT desktop.

codex plugin marketplace add Lycatra/lycatra-plugin
codex plugin add lycatra@lycatra

Codex IDE extension (tools only):

codex mcp add lycatra --url https://lycatra.com/api/mcp
Cursor

Lycatra tools in Cursor's agent.

VS Code

GitHub Copilot in VS Code.

GitHub Copilot CLI

Lycatra in the Copilot command line.

copilot plugin marketplace add Lycatra/lycatra-plugin
copilot plugin install lycatra@lycatra
Gemini CLI

Lycatra tools as a Gemini CLI extension.

gemini extensions install https://github.com/Lycatra/lycatra-plugin --auto-update
Goose

Lycatra tools as a Goose extension.

LM Studio

Lycatra tools for local models.

Any other MCP client

Add the MCP server URL. It signs in with OAuth.

Set up every agent add-mcp finds:

npx add-mcp https://lycatra.com/api/mcp
The Lycatra CLI

Agents that can run shell commands install it themselves the first time they need your computer, for example to run a command inside one of your agents. To install it yourself:

curl -fsSL https://lycatra.com/install.sh | bash
irm https://lycatra.com/install.ps1 | iex

How it works

Every app talks to Lycatra's hosted MCP server for anything stored in your account: agents, wiki, memory, schedules, notifications, teams, Matrix, vault metadata, and a browser that runs inside your agents. You sign in once with Lycatra's OAuth prompt, and the app only sees what your account can access.

Hosted MCP and the Lycatra CLI are both supported interfaces. An authorized local or cloud shell can use the checksum-verified CLI for the same account operations. Local ports and child processes additionally need suitable runtime capabilities. Ordinary hosted tools do not require a personal computer.

Cloud event delivery is negotiated with the receiving host when supported by the server. It continues independently of a CLI shell. The CLI can inspect and manage owner-authorized subscriptions, but cannot manufacture a host callback or wake a chat by name. Start with lycatra notify targets doctor --json for read-only diagnosis. Local notify wake adapters remain separate.

For explicit cloud-agent enrollment, use the selected environment's /api/mcp/cloud resource returned by notify targets connect. Compatible hosts discover its starter scopes and follow OAuth enrollment and consent. The normal /api/mcp resource retains human-account behavior. Cloud event support still requires qualification in the receiving host.

Explicitly approved notify events subscriptions resync <id> --yes skips one parked or cancelled head event and reports the skipped sequence and new cursor. It does not retry that event or change subscription pause or lease expiry.

Beta uses dev and stable uses production throughout installation, login and updates. Each CLI profile has separate credentials and update state. Provider credential setup for local coding harnesses is optional, separate from login.

Your app's normal permission prompts still apply to writes, deletions, and anything visible to other people.

Support

Comments

Markdown supported

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.