AUTOMATIONSWITCH
OfficialAI / ML

Weights & Biases MCP

by Weights & Biases

Official W&B MCP server for Weights & Biases Models and Weave. Query experiments, runs, sweeps, models, traces, evaluations through MCP. 50 GitHub stars and 13 commits on main in the last 30 days.

50·7 tools·Released JAN 2026·MIT
pip install wandb-mcp-server
Share:

Official W&B MCP server for Weights & Biases Models and Weave. Query experiments, runs, sweeps, models, traces, evaluations through MCP. 50 GitHub stars and 13 commits on main in the last 30 days.

Reviewed by M. Nouriel · MAY 2026

INSTALL THIS SERVER

Requires authenticationWANDB_API_KEY environment variable. Token from wandb.ai user settings.
{ "mcpServers": { "wandb": { "command": "python", "args": [ "-m", "wandb_mcp_server" ], "env": { "WANDB_API_KEY": "<your-wandb-api-key>" } } } }
PrereqRequires WANDB_API_KEY from wandb.ai user settings. PyPI: `wandb-mcp-server`. Weave tools require the W&B Weave product. Path: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS).
{ "mcpServers": { "wandb": { "command": "python", "args": [ "-m", "wandb_mcp_server" ], "env": { "WANDB_API_KEY": "<your-wandb-api-key>" } } } }
{ "mcpServers": { "wandb": { "command": "python", "args": [ "-m", "wandb_mcp_server" ], "env": { "WANDB_API_KEY": "<your-wandb-api-key>" } } } }
{ "mcpServers": { "wandb": { "command": "python", "args": [ "-m", "wandb_mcp_server" ], "env": { "WANDB_API_KEY": "<your-wandb-api-key>" } } } }
{ "mcpServers": { "wandb": { "command": "python", "args": [ "-m", "wandb_mcp_server" ], "env": { "WANDB_API_KEY": "<your-wandb-api-key>" } } } }

7 TOOLS AVAILABLE

list_runs
List W&B runs in a project
Read
get_run
Get details for a specific run
Read
query_metrics
Query run metrics over a time range
Read
list_artifacts
List artifacts (datasets, models) for a project
Read
list_sweeps
List hyperparameter sweeps
Read
query_weave_traces
Query Weave LLM traces
Read

OUR ASSESSMENT

Strengths
  • Official W&B maintenance.
  • 50 GitHub stars and MIT licence.
  • 13 commits on main in the last 30 days.
  • Covers both W&B Models (experiment tracking) and W&B Weave (LLM observability) in one MCP.
  • Useful as a memory and lookup layer for agents that need to reference past experiments or evaluations.
Weaknesses
  • 50 GitHub stars; adoption signal is early.
  • Weave coverage requires the W&B Weave product; teams using W&B Models only see a subset of value.
  • W&B API key grants account-scoped access; rotate via W&B settings.
Security Notes

W&B API key is account-scoped: the MCP sees what the key holder sees. Use a dedicated service account API key for production agents. Weave traces can include LLM input and output; treat the result stream as sensitive when the underlying training or evaluation data is sensitive.

Best For

ML teams running W&B for experiment tracking who want agents to look up past runs and metrics; LLM engineering teams using W&B Weave for trace observability who want agent access to traces and evaluations; agentic ML workflows that need to reference experiment history as context.

TECHNICAL DETAILS

Language
python
Transport
stdio
Clients
Claude DesktopClaude CodeCursorVS CodeWindsurf
License
MIT
GitHub
npm
wandb-mcp-server
Last Release
wandb-mcp-server (PyPI latest)MAY 3, 2026
First Released
JAN 1, 2026

ADOPTION METRICS

// GitHub Stars
50

// Reading this50 stars on the wandb/wandb-mcp-server repo. 13 commits on main in the last 30 days. Official W&B maintenance carries the editorial weight.

// Popularity Rank
#10
Globally · #10 in AI / ML

// Reading thisPairs with Phoenix (LLM observability) in the ai-ml category for ML experiment tracking and trace inspection.

SOURCES & VERIFICATION

We don't take any single directory's word for it. Before scoring, we cross-reference 4 public MCP sources, install the server ourselves against the clients we cover, and record when we last re-verified.

01
Discovered
Manual submission
First indexed MAY 3, 2026
02
Cross-referenced
4 directories
PulseMCP, MCP.so, Glama, Official MCP Registry
03
Verified against
Claude Desktop, Cursor
Installed and tested across clients
04
Last re-checked
MAY 3, 2026
Weekly re-verification
// How other directories see it

The same server, 4 different lenses. We reconcile these signals into our editorial score, which is why our number sometimes diverges from a directory-aggregate star count.

SourceTheir ratingTheir star countTheir downloadsLast synced
AutomationSwitch This page4.2editorial50MAY 3, 2026
PulseMCP— unratedunavailableunavailableMAY 3, 2026
MCP.so— unratedunavailableunavailableMAY 3, 2026
Glama— unratedunavailableunavailableMAY 3, 2026
Official MCP Registry— unratedunavailableunavailableMAY 3, 2026

// Counts are directory-reported; we don't adjust them. Discrepancies usually come from different snapshot times or star-caching.

OTHER AI / ML MCP SERVERS

Community4.6

Cognee MCP

topoteretes

Knowledge graph plus vector memory engine for AI agents, exposed as an MCP server with V2 session-aware memory tools (remember, recall, forget, improve) and classic V1 ingestion pipelines (cognify, codify). Three transports: stdio, SSE, Streamable HTTP. 16,965 GitHub stars, Apache-2.0.

8 tools16,997
Community4.6

Codebase Memory MCP

DeusData

High-performance code intelligence MCP server for AI coding agents. Indexes a codebase into a queryable knowledge graph in milliseconds, with 14 MCP tools spanning structural search, call-chain tracing, impact analysis, dead-code detection, and Cypher queries. Single static C binary, 66 languages via tree-sitter, zero runtime dependencies.

9 tools2,021
Vendor4.5

Arize Phoenix MCP

Arize AI

LLM observability platform exposing prompts, projects, traces, spans, sessions, datasets, and experiments through MCP. Published to npm as @arizeai/phoenix-mcp, current 4.0.8 (2026-04-29). 9,496 stars on parent monorepo, Elastic License 2.0.

8 tools9,496
Vendor4.5

Qdrant MCP Server

Qdrant

Official Qdrant vector database MCP server. Acts as a semantic memory layer on top of Qdrant: store information with metadata, retrieve via similarity search. Two tools, very small surface area, exceptionally maintained by the Qdrant team. Configurable embedding provider (fastembed default), supports remote and local Qdrant clusters.

2 tools1,373
Official4.4

Amazon Bedrock AgentCore MCP

AWS Labs

Official AWS Labs MCP server for Amazon Bedrock AgentCore: agent runtime, memory, gateway, identity, and observability. Tools fetch curated AgentCore documentation and surface deployment guides for runtime, memory, and gateway resources. Apache-2.0 within awslabs/mcp monorepo (8,924 parent stars).

5 tools8,924
Official4.1

Amazon Bedrock Knowledge Base MCP

AWS Labs

Official AWS Labs MCP for Bedrock Knowledge Base retrieval: discover knowledge bases, query with natural language, filter by data source, and rerank results. Apache-2.0 within awslabs/mcp monorepo. Tight tool surface focused on RAG over AWS-managed KBs.

4 tools8,924
// Get in touch

DISCUSS YOUR
MCP REQUIREMENTS.

Evaluating a server, scoping an internal deployment, or working out whether MCP is the right fit at all. Start the conversation and we will point you at the right piece of the ecosystem.

Discuss Your MCP Requirements →