datalens-mcp

MCP server for the Yandex DataLens public API: 70 tools for dashboards, charts, datasets and connections

  • 8 ★ on the repository
  • Last commit
  • MIT
MCP server

Extended access

Why this level

  • Requires a Yandex Cloud IAM token
  • Tools create and modify objects, including access bindings and public embeds
  • Large responses with connection config are saved to temp files
All reasons and checks
Russian stack

osipsemyon/datalens-mcp

Install

Manual install

npm install
npm run build
claude mcp add datalens -- node /path/to/datalens-mcp/dist/index.js

Run inside a clone of the repo, then set the environment variables.

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • Requires a Yandex Cloud IAM token
  • Tools create and modify objects, including access bindings and public embeds
  • Large responses with connection config are saved to temp files

This is third-party code. Review the repository files before installing.

What it does

Wraps nearly the whole Yandex DataLens public API as 70 MCP tools generated from the official OpenAPI spec. The agent can read and create dashboards, charts, datasets, connections, workbooks and access bindings. Delete tools are intentionally omitted, and DATALENS_READ_ONLY=1 keeps only read tools. The composite clone_entry tool copies objects server-side.

Who it is for. Analysts and BI developers who build reporting in Yandex DataLens.

Good fit when

  • Understand the structure of an existing workbook, dashboard or dataset
  • Clone a dashboard or chart and point it at a new dataset
  • Audit access bindings and connections
  • Move objects between workbooks in bulk

Not a fit when

  • You need to delete objects: the server cannot
  • You need to edit an existing dataset in place: the DataLens API currently rejects such edits

Example request

List all dashboards in the sales workbook and show which datasets their charts use

MCP

Transport
stdio
Authentication
API key

Environment variables

DATALENS_IAM_TOKEN required, secret
IAM token from yc iam create-token, valid for about 12 hours
DATALENS_ORG_ID required
DataLens organization ID for the x-dl-org-id header
DATALENS_READ_ONLY
1 registers read tools only
DATALENS_USE_YC_TOKEN
1 mints a token via yc when using datalens-launch.sh
DATALENS_API_BASE
API base URL for self-hosted DataLens

Limitations

Needs an IAM token that lasts about 12 hours and is read once at startup, so restart the server after refreshing it. Yandex OAuth tokens are not supported. Editing existing datasets fails with DATASET_REVISION_MISMATCH on the API side; use clone_entry instead. DataLens marks dashboard and chart create and update as experimental. The confirmation policy is advisory; only read-only mode is a hard guarantee.

How to disable. Run claude mcp remove datalens or remove the datalens block from ~/.claude.json.

FAQ

How do I avoid refreshing the IAM token by hand?

Launch via bin/datalens-launch.sh with DATALENS_USE_YC_TOKEN=1: it mints a fresh token with the yc CLI on every start.

What about very large responses?

If a result exceeds DATALENS_MAX_RESPONSE_CHARS, the full JSON is written to a temp file and the tool returns its path.

Does it work with self-hosted DataLens?

Yes, set DATALENS_API_BASE; it must be https, with http allowed for localhost or with DATALENS_ALLOW_INSECURE_BASE=1.

Discussion

No questions or reviews yet.

DataLens Dev MCP

datalens-dev-mcp

Russian stack

Local MCP server and Codex plugin for Yandex DataLens dashboard development: read, author, publish, back up and clean up objects

MCP serverBroad access7Repository stars
Official

The official DataLens MCP server from the development team: turns the public DataLens API into three gateway tools instead of hundreds of separate MCP tools

MCP serverExtended access4Repository stars
Russian stack

A Rust MCP server for the Yandex DataLens public API: 60 methods via a full catalog, typed wrappers for common operations and IAM auth

MCP serverExtended access5Repository stars
Russian stack

A translator from the cloud DataLens API to internal open-source DataLens 2.9.0 calls, plus a Streamable HTTP MCP server on top

MCP serverExtended access5Repository stars
Official

Official Yandex demo templates: MCP gateways over cloud functions for YandexART, charts, PDF, arXiv, Kaggle and weather

MCP serverExtended access1Repository stars
Official

Yandex Cloud's official MCP server set: documentation, search, app deployment, serverless functions, containers and a data catalog

MCP serverBroad access49Repository stars

README in short

The server runs over stdio and splits 70 tools into read, create and modify kinds with different confirmation levels. Tools are generated from the OpenAPI spec and checked by a conformance audit. The README covers IAM token auth, token auto-refresh via yc, large response handling and Claude Code setup. DataLens API limits for datasets and experimental objects are described separately.

README badge

Report a problem

Foxx AIdatalens-mcp
Foxx AI

I am Foxx AI and I have already vetted this tool. Ask about install, setup or anything else, and I will keep it simple.