Colab MCP Server

Source-backed facts for citing this resource, derived directly from the registry — also available as plain text for AI assistants.

Canonical URL

https://heyclau.de/entry/mcp/colab-mcp-server

Source URLs

https://raw.githubusercontent.com/googlecolab/colab-mcp/main/README.md, https://github.com/googlecolab/colab-mcp

Brand

Colab MCP

Brand domain

colab.research.google.com

Brand asset source

brandfetch

Safety notes

Colab MCP runs a local websocket proxy and waits for a Google Colab browser session to connect before proxying session tools., The server creates a bearer-style proxy token and accepts a single authorized Colab-origin websocket connection at a time., A connected Colab runtime can execute code and interact with notebooks, outputs, files, variables, packages, and mounted resources depending on the session and tools exposed by Colab., Use disposable notebooks or reviewed development sessions before allowing Claude to run code or inspect outputs., Avoid connecting notebooks that have access to production data, cloud credentials, private datasets, paid accelerators, or long-running jobs unless the workflow has explicit approval.

Privacy notes

Notebook code, prompts, outputs, logs, variables, datasets, file paths, runtime metadata, package lists, browser session details, and generated websocket tokens can be visible to the MCP client and model provider., Colab notebooks may contain API keys, OAuth tokens, mounted Drive paths, secrets in environment variables, private model weights, customer data, or unpublished research., Local log files are created under a temporary Colab MCP log directory by default unless a log directory is specified., Review Google Colab, MCP client, model provider, and organization retention policies before sending notebook or runtime context to an assistant.

Author

Google Colab

Submitted by

oktofeesh1

Claim status

unclaimed

Last verified

2026-06-06

• Notebook code, prompts, outputs, logs, variables, datasets, file paths, runtime metadata, package lists, browser session details, and generated websocket tokens can be visible to the MCP client and model provider.
• Colab notebooks may contain API keys, OAuth tokens, mounted Drive paths, secrets in environment variables, private model weights, customer data, or unpublished research.
• Local log files are created under a temporary Colab MCP log directory by default unless a log directory is specified.
• Review Google Colab, MCP client, model provider, and organization retention policies before sending notebook or runtime context to an assistant.

• uv or uvx available to run the server from the GitHub repository.
• Local MCP client that supports `notifications/tools/list_changed`.
• Google Colab account and browser session you are authorized to use.
• Local machine access, because upstream documents that the MCP client must run locally.
• Review of notebook, runtime, file, credential, and data access before connecting an agent to a Colab session.

Show that Colab MCP Server is listed on HeyClaude. Paste this Markdown into your README — it renders the badge and links back to this page.

[![Listed on HeyClaude](https://heyclau.de/badge/mcp/colab-mcp-server.svg)](https://heyclau.de/entry/mcp/colab-mcp-server)

Colab MCP Server side by side with 3 alternatives on trust, install, platform support, and disclosed safety notes — all from reviewed registry metadata.

Field	Colab MCP Server Connect Claude to a Google Colab runtime. Colab MCP starts a local FastMCP server and websocket proxy, waits for an authorized Colab-origin browser connection, then exposes the session's notebook and runtime tools to your MCP client over stdio. Open dossier	Jupyter MCP Server MCP server for connecting Claude to Jupyter notebooks, kernels, files, cells, multimodal outputs, and JupyterLab workflows over stdio or Streamable HTTP. Open dossier	AI Game Developer Unity MCP Server Unity MCP server, plugin, CLI, and skill generator for controlling Unity Editor and runtime projects from MCP clients through built-in game-dev tools. Open dossier	Django MCP Server Django extension that exposes MCP endpoints and stdio transport for Django apps, with declarative model query tools, custom toolsets, DRF create/list/ update/delete tool publishing, serializer output, and MCP inspection. Open dossier
Next steps	Open dossierCopy installCopy configAPI JSONOpen LLMMCP feedOpen sourceNewsletterClaim listing	Open dossierCopy installCopy configAPI JSONOpen LLMMCP feedOpen sourceNewsletterClaim listing	Open dossierCopy installCopy configAPI JSONOpen LLMMCP feedOpen sourceNewsletterClaim listing	Open dossierCopy installCopy configAPI JSONOpen LLMMCP feedOpen sourceNewsletterClaim listing
Trust	Review first	Review first	Review first	Review first
Review status	Not reviewed	Not reviewed	Not reviewed	Not reviewed
Package trust	Package not verified	Package not verified	Package not verified	Package not verified
Source provenance	Source-backed	Source-backed	Source-backed	Source-backed
Submitter	oktofeesh1	oktofeesh1	oktofeesh1	oktofeesh1
Install risk	Review first	Review first	Review first	Review first
Notes	Safety ✓ Privacy ✓	Safety ✓ Privacy ✓	Safety ✓ Privacy ✓	Safety ✓ Privacy ✓
Brand	Colab MCP	Jupyter MCP Server	AI Game Developer	Django MCP Server
Category	mcp	mcp	mcp	mcp
Source	Source-backed	Source-backed	Source-backed	Source-backed
Author	Google Colab	Datalayer	Ivan Murzak	Smart GTS
Added	2026-06-06	2026-06-06	2026-06-06	2026-06-06
Platforms	Claude CodeCodexCursorClaude Desktop	Claude CodeClaude Desktop	Claude CodeClaude Desktop	Claude CodeClaude Desktop
Harness	Claude CodeCodexCursorClaude Desktop	Claude CodeClaude Desktop	Claude CodeClaude Desktop	Claude CodeClaude Desktop
Source repo	—	—	—	—
Safety notes	✓Colab MCP runs a local websocket proxy and waits for a Google Colab browser session to connect before proxying session tools. The server creates a bearer-style proxy token and accepts a single authorized Colab-origin websocket connection at a time. A connected Colab runtime can execute code and interact with notebooks, outputs, files, variables, packages, and mounted resources depending on the session and tools exposed by Colab. Use disposable notebooks or reviewed development sessions before allowing Claude to run code or inspect outputs. Avoid connecting notebooks that have access to production data, cloud credentials, private datasets, paid accelerators, or long-running jobs unless the workflow has explicit approval.	✓Jupyter MCP Server can list files, list kernels, create or switch notebooks, insert cells, edit cell source, delete cells, move cells, restart kernels, execute cells, and execute arbitrary code in a notebook kernel. Notebook code can read files available to the Jupyter process, access environment variables, call network services, start subprocesses, install packages, or mutate data depending on kernel permissions. File and notebook tools can alter important notebooks or delete work if pointed at the wrong server, path, notebook, or cell index. Keep Claude connected only to isolated kernels and directories that are appropriate for agent-driven execution. Use least-privilege Jupyter tokens, short-lived sessions, and a separate test notebook before allowing changes in production or shared JupyterHub environments. Streamable HTTP deployments require MCP client authentication unless explicitly started in insecure no-auth mode; do not expose no-auth transports beyond trusted local testing.	✓AI Game Developer can create, move, copy, modify, and delete Unity assets, scenes, GameObjects, components, scripts, packages, and generated project files. Tools include dynamic C# script execution, C# reflection method lookup and calls, package installation/removal, Unity test execution, editor state changes, play mode control, screenshots, and profiler access. The MCP server supports streamable HTTP and stdio; HTTP deployments should require a bearer token and stay bound to trusted interfaces. Server variables include optional authorization and webhook settings; webhook endpoints can receive tool, prompt, resource, connection, and authorization events. Unity runtime connections can expose compiled game state or in-game behavior to an MCP client, not just editor-only project data. Use source control, backups, tool filtering, and explicit review before allowing destructive tools such as asset deletion, package removal, script execution, or reflection calls.	✓Django MCP Server can expose Django model querysets, custom Python methods, DRF create/list/update/delete views, serializers, resources, and low-level FastMCP tools to an MCP client. Published DRF create, update, and delete tools can mutate application data if their serializers, views, and authentication rules permit it. The README notes that built-in DRF authentication classes, permission classes, filter backends, and pagination are disabled for published DRF tools in favor of MCP authentication; review this carefully before reusing production views. Query tools can evaluate QuerySets and return database records; restrict queryset scope and fields before exposing sensitive models. Require confirmation and application-level authorization before exposing write tools, email-sending methods, admin-like actions, or tools that touch customer, employee, financial, health, or regulated data.
Privacy notes	✓Notebook code, prompts, outputs, logs, variables, datasets, file paths, runtime metadata, package lists, browser session details, and generated websocket tokens can be visible to the MCP client and model provider. Colab notebooks may contain API keys, OAuth tokens, mounted Drive paths, secrets in environment variables, private model weights, customer data, or unpublished research. Local log files are created under a temporary Colab MCP log directory by default unless a log directory is specified. Review Google Colab, MCP client, model provider, and organization retention policies before sending notebook or runtime context to an assistant.	✓Jupyter tokens, notebook contents, code cells, markdown cells, outputs, plots, image data, file names, kernel lists, environment-derived values, and error traces can be exposed to the MCP client. Executed notebook cells may print secrets, local paths, database results, API responses, or private research data into model context or client logs. Multimodal output support can send generated plots and images to the MCP client; disable or scope `ALLOW_IMG_OUTPUT` when images may contain sensitive data. Shared notebooks and JupyterHub deployments may contain other users' work; confirm workspace boundaries before connecting an agent. Retain notebook execution logs, MCP logs, and generated outputs only as long as needed for the workflow.	✓Tool calls may expose Unity project paths, asset names, scene hierarchy, serialized object data, scripts, logs, screenshots, profiler metrics, test output, package metadata, and runtime state. MCP config files and Unity plugin config can contain server URLs, connection modes, bearer tokens, authorization settings, enabled tool IDs, and cloud or local endpoint details. Optional webhooks can receive tool, prompt, resource, connection, authorization, and token-bearing request data. Screenshots from Game View, Scene View, cameras, or isolated GameObjects may include proprietary artwork, level design, UI, debug overlays, or unreleased game content. Generated skills and AI-client configuration files can reveal available tools, project structure, installed packages, and local workflow assumptions.	✓Django sessions, request headers, model names, field names, primary keys, QuerySet results, serializer output, DRF request bodies, custom tool arguments, and tool responses can be exposed to the MCP client. Exposed models may contain user accounts, permissions, customer records, orders, messages, files, logs, internal notes, audit trails, or application-specific secrets. Remote streamable HTTP deployments can move application data outside the original Django UI and audit path if MCP auth, OAuth metadata, and retention are not configured correctly. Stdio usage can still expose data through local MCP client logs, transcripts, and tool traces. Keep MCP endpoint access, serializer fields, queryset filters, and tool docstrings intentionally narrow for each app.
Prerequisites	uv or uvx available to run the server from the GitHub repository. Local MCP client that supports `notifications/tools/list_changed`. Google Colab account and browser session you are authorized to use. Local machine access, because upstream documents that the MCP client must run locally.	Python 3.10 or newer with `uvx` available. A running JupyterLab, Jupyter Server, JupyterHub, or compatible notebook deployment. A Jupyter access token for the notebook server Claude is allowed to control. Review of notebook trust, kernel permissions, filesystem scope, and network access before enabling code-execution tools.	Unity project using Unity 2022.3 or newer for the current package metadata. Unity Hub or a local Unity Editor installation. Node.js 20.19 or newer, or Node.js 22.12 or newer, for `unity-mcp-cli`. An MCP client such as Claude Code, Claude Desktop, Codex, Cursor, Gemini CLI, GitHub Copilot, Cline, or another supported client.	Django 4 or 5 application with Python 3.10 or newer. mcp_server added to INSTALLED_APPS and mcp_server.urls included in the Django URL configuration. Review of which Django models, querysets, custom methods, DRF views, serializers, and request context should be exposed to MCP clients. Authentication classes configured through DJANGO_MCP_AUTHENTICATION_CLASSES before exposing non-public data over streamable HTTP.
Install	uvx git+https://github.com/googlecolab/colab-mcp Copy install	uvx jupyter-mcp-server@latest Copy install	npm install -g unity-mcp-cli && unity-mcp-cli install-plugin ./MyUnityProject Copy install	pip install django-mcp-server Copy install
Config	{ "mcpServers": { "colab-mcp": { "command": "uvx", "args": ["git+https://github.com/googlecolab/colab-mcp"], "timeout": 30000 } } } Copy config	Manual-only setup: claude mcp add jupyter --env JUPYTER_URL=YOUR_JUPYTER_SERVER_URL --env JUPYTER_TOKEN=YOUR_JUPYTER_TOKEN -- uvx jupyter-mcp-server@latest Copy config	Manual-only setup: npm install -g unity-mcp-cli unity-mcp-cli install-plugin ./MyUnityProject unity-mcp-cli open ./MyUnityProject Copy config	{ "mcpServers": { "django": { "command": "python", "args": [ "manage.py", "stdio_server" ], "type": "stdio" } } } Copy config
Citations	Source repositorygithub.com 2026-07-21T07:17:01+00:00 Documentationraw.githubusercontent.com Submitted by oktofeesh12026-06-06 Source methodology →	Source repositorygithub.com 2026-07-21T07:17:01+00:00 Documentationraw.githubusercontent.com Submitted by oktofeesh12026-06-06 Source methodology →	Source repositorygithub.com 2026-07-21T07:17:01+00:00 Documentationraw.githubusercontent.com Submitted by oktofeesh12026-06-06 Source methodology →	Source repositorygithub.com 2026-07-21T07:17:01+00:00 Documentationraw.githubusercontent.com Submitted by oktofeesh12026-06-06 Source methodology →
Claim	Unclaimed	Unclaimed	Unclaimed	Unclaimed

Source-backed guides for putting this to work.