MCP: FetchSandbox MCP
by FetchSandbox
A sandbox for third-party API integrations: agents replay Stripe, GitHub, Twilio and 50+ other workflows — webhooks, retries and failure cases included — without touching the real API.
Drive live Jupyter notebooks from an agent — read and edit cells, execute code, see rich outputs, and run against local JupyterLab or a cloud sandbox.
Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.
Datalayer's Jupyter MCP Server connects an agent to a running Jupyter kernel rather than to notebook files on disk. That distinction is the whole point: the agent can insert and edit cells, execute them, read the outputs — including images, when ALLOW_IMG_OUTPUT is enabled — and iterate the way a person would, with state carried between cells instead of re-running a script from scratch each time.
It handles multiple notebooks in one session, exposes server and sandbox management tools alongside cell operations, and ships MCP prompts as well as tools. A JupyterLab integration adds further tools that can be switched on at the command line.
Where the kernel lives is configurable. Beyond a local JupyterLab, sandbox variants run the code on Datalayer, Kaggle (batch or interactive, with optional GPU), Google Colab or Modal, so the same agent workflow scales from a laptop to accelerated cloud compute. A hosted version of the server is also available, and there is a Claude Code plugin (/plugin marketplace add datalayer/jupyter-mcp-server).
Setup is a short list: pip install jupyterlab jupyter-collaboration jupyter-mcp-tools ipykernel, start JupyterLab with a token, then point the client at it. Distributed under the BSD 3-Clause licence.
Covered in the Weekly
by FetchSandbox
A sandbox for third-party API integrations: agents replay Stripe, GitHub, Twilio and 50+ other workflows — webhooks, retries and failure cases included — without touching the real API.
by Daytona
Daytona's official MCP server: give an agent a disposable cloud sandbox with a file system, git, process execution, computer use and a live preview URL.
by Plaid
Plaid's official sandbox MCP server: search Plaid docs, mint sandbox access tokens, generate mock financial data and fire test webhooks while building an integration.
Context-compression MCP server that shrinks tool outputs, logs, files and RAG chunks before they reach the model, cutting 60-95% of tokens on JSON payloads.