Amika MCP Server

Amika MCP Server

Remote MCP server for Amika: manage sandboxes and agent sessions from any MCP client, including Claude mobile app.

Category
Visit Server

README

Amika MCP Server

Deploy to Cloudflare

A remote MCP server for Amika: manage your sandboxes and agent sessions from any MCP client — including the Claude mobile app, so you can check on and steer your coding agents from your phone.

You deploy your own instance to Cloudflare Workers (free tier is plenty) with your own Amika API key. Nothing is shared: your key stays in your worker.

Tools

Tool What it does
list_sandboxes List all sandboxes with their state
get_sandbox Details/state of one sandbox
start_sandbox / stop_sandbox Start or stop a sandbox (waits for the state change)
list_sessions List agent sessions in a sandbox
get_latest_session Most recent session in a sandbox
get_session Fetch one session by id
create_session Create an empty session with optional metadata
update_session Update a session's status/metadata (e.g. mark completed)
send_to_agent Send a message to the Claude/Codex agent in a sandbox and wait for the reply — resumes the latest session by default, or pass new_session / session_id

Deploy

One-click

  1. Click Deploy to Cloudflare above. Cloudflare copies this repo into your GitHub account and creates a Worker that redeploys on every push.
  2. Add two secrets to the worker (dashboard → your worker → Settings → Variables and Secrets, type Secret):
    • AMIKA_API_KEY — generate at app.amika.dev/settings (API Key section)
    • MCP_PATH_SECRET — a long random string, e.g. from openssl rand -hex 24

Your MCP endpoint is:

https://<worker-name>.<your-subdomain>.workers.dev/<MCP_PATH_SECRET>/mcp

CLI

git clone https://github.com/Neolithic/amika-mcp
cd amika-mcp && npm install
npx wrangler login                       # or: export CLOUDFLARE_API_TOKEN=<token>
npx wrangler secret put AMIKA_API_KEY
npx wrangler secret put MCP_PATH_SECRET
npm run deploy

If wrangler login's browser flow can't reach you (headless/remote machine), create an API token with the "Edit Cloudflare Workers" template at dash.cloudflare.com/profile/api-tokens and set CLOUDFLARE_API_TOKEN instead.

Connect a client

Claude (including the mobile app) — requires a Pro/Max/Team plan:

  1. claude.ai/settings/connectors → Add custom connector → name it Amika, paste your full endpoint URL (including the secret segment). Leave the OAuth fields empty.
  2. In any Claude app (web, desktop, mobile), enable the connector in the tools menu and ask things like "list my amika sandboxes" or "send 'run the tests' to the agent in dev-box".

Claude Code:

claude mcp add --transport http amika https://<worker>/<secret>/mcp

Any other MCP client that speaks Streamable HTTP works the same way.

Security model

  • Single-tenant by design. Each deployment serves exactly one Amika account. The worker holds your AMIKA_API_KEY as a Cloudflare secret; it is never sent to clients.
  • The URL is the credential. There is no OAuth — the endpoint is guarded by the secret path segment. Treat the full URL like a password: don't share it, don't post screenshots of it, and rotate it if it leaks (wrangler secret put MCP_PATH_SECRET, or edit it in the dashboard).
  • Never share your deployed endpoint with other people. Anyone with the URL has full control of your Amika account's sandboxes and agents. Other users should deploy their own instance — that's what the button is for.
  • Requests with a wrong or missing path secret get a 404 without touching the Amika API.

Local development

cp .dev.vars.example .dev.vars   # fill in AMIKA_API_KEY and MCP_PATH_SECRET
npm run dev                      # serves http://localhost:8787/<secret>/mcp
npm run typecheck

Notes & limitations

  • send_to_agent blocks until the agent finishes (Amika's endpoint is synchronous with a 10-minute timeout). Long agent runs may exceed your MCP client's own tool timeout — the work still continues inside Amika; check back with get_latest_session.
  • The Amika SDK stores an unbound fetch reference, which throws "Illegal invocation" on Workers. The client passes fetch: (input, init) => fetch(input, init) to work around it (src/index.ts) — don't remove it.
  • Built with @amika/sdk, the MCP TypeScript SDK, and Cloudflare's agents framework (Streamable HTTP via McpAgent).

License

MIT

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured