MCP Tools DirectoryAbout this integration
Execute Python code on Google Colab GPU runtimes (T4/L4) from any MCP client
- Transport
- stdio
- Authentication
- oauth
- Initial setup
- user-authorization
- Runtime
- unattended
- Evidence
- documented
- Version
- 0.1.0
- Package
- mcp-server-colab-exec
- Last compatibility test
- Not independently tested
Connect your agent
mcp-server-colab-execPublisher documentation reviewed offline; package, OAuth flow, and Colab runtime were not executed. This server executes supplied code remotely and may consume Colab quota; it should be enabled only with an appropriate execution and artifact-handling policy.
Capabilities: Execute inline Python on a Colab GPU runtime, Execute workspace Python files on a Colab GPU runtime, Collect generated notebook artifacts
Connected profiles
Additional details
io.github.pdwi2020/mcp-server-colab-exec
Source ↗ · Checked 2026-09-170.1.0
Source ↗ · Checked 2026-09-17CC0-1.0; package licenses are separate
Source ↗ · Checked 2026-09-17pypi
Source ↗ · Checked 2026-09-17mcp-server-colab-exec
Source ↗ · Checked 2026-09-190.1.0
Source ↗ · Checked 2026-09-19The first run requires browser consent; subsequent server runs use the cached token and refresh it automatically, allowing configured MCP calls without repeating browser consent each time.
Source ↗ · Checked 2026-09-19Uses Google OAuth2 user authorization and caches access and refresh tokens for subsequent runs.
Source ↗ · Checked 2026-09-19The MCP host launches the `mcp-server-colab-exec` console script as a local MCP subprocess.
Source ↗ · Checked 2026-09-19Publisher documentation reviewed; package and integration endpoint not executed or independently security-audited.
Source ↗ · Checked 2026-09-19