Which integrations are available for this server?

Allows connection to Databricks-hosted MLflow tracking servers using Databricks PAT or service-principal tokens, enabling management of experiments, runs, models, and traces on Databricks. Provides comprehensive tools for managing MLflow experiments, runs, registered models, model versions, logged models (MLflow 3), traces, assessments, webhooks, and prompt-optimization jobs. Includes aggregation tools and MCP resources for efficient AI-driven workflows.

de en es ja ko ru zh

mlflow-mcp-server

by us-all

Overview Schema Source Related Servers Discussions Admin

TypeScript

Remote

The MLflow MCP Server provides 82 tools across the full MLflow lifecycle — experiment tracking, model registry, tracing, assessments, webhooks, and prompt optimization — accessible via the Model Context Protocol.

Experiment Management: Create, search, get, update, delete, restore experiments; set/delete tags; look up by name or ID.

Run Tracking: Create, get, search, update, delete, restore runs; log metrics, params, and tags individually or in batch; log dataset inputs; retrieve metric history; list artifacts; find best run by metric; compare runs side-by-side; search runs by tag.

Model Registry: Create, get, search, rename, update, delete registered models; manage model versions (create, get, search, update, delete, transition stages, get download URI); set/delete tags and aliases (e.g. champion).

MLflow 3 Logged Models: Create, search, get, finalize, delete LoggedModel entities; set/delete tags; log parameters.

Trace Management: Search/filter traces, retrieve full details or metadata-only, delete traces, set/delete tags; list and get trace attachments (Databricks only).

Assessments: Log feedback (scores, labels, judgments) and ground-truth expectations on traces; get, update, delete assessments.

Webhooks: Create, list, get, update, delete, and test webhooks for model registry events (version creation, stage transitions, alias/tag changes).

Prompt Optimization: Create, get, search, cancel, delete prompt optimization jobs to automatically improve registered prompts.

Aggregation Tools: summarize-experiment (overview + top-N runs + metric stats in one call) and summarize-run (run info + metric history + artifacts in one call).

MCP Prompts: Four workflow templates — debug-failed-traces, promote-best-run, compare-top-runs, and annotate-trace-quality.

MCP Resources: URI-based access to entities (e.g. mlflow://run/{runId}, mlflow://registered-model/{name}).

Efficiency & Control:

search-tools for natural-language tool discovery
extractFields projection to reduce response token size
Category-based tool filtering via MLFLOW_TOOLS / MLFLOW_DISABLE env vars
Read-only by default; writes require MLFLOW_ALLOW_WRITE=true
Supports stdio and Streamable HTTP transports with Bearer/Basic auth

MLflow MCP Server

The widest-coverage MLflow MCP — including MLflow 3 traces, prompt-optimization, webhooks, and Databricks trace attachments that no other MCP exposes.
82 tools across experiments, runs, registry, logged models, traces, assessments, webhooks, prompt-optimization. Aggregation tools (summarize-experiment, summarize-run) fold 3–5 round-trips into one structured response with already-fetched metric stats.

What it does that others don't

Full coverage — only third-party MLflow MCP shipping prompt-optimization-jobs (5 tools), webhooks (6), MLflow 3 LoggedModel (8), and Databricks trace attachments (list-trace-attachments, get-trace-attachment — Databricks MLflow only; OSS returns 404).
Aggregation tools — summarize-experiment returns experiment + topN runs + metric stats (min/max/mean) in one call from already-fetched data, zero extra round-trips. summarize-run dedups metricHistory.history.*.key (~100KB savings on 4k-point series).
MCP Prompts (4) — debug-failed-traces, promote-best-run, compare-top-runs, annotate-trace-quality. Workflow templates the model invokes directly.
MCP Resources (6) — mlflow://run/{runId}, mlflow://experiment/{expId}, mlflow://run/{runId}/artifacts, mlflow://experiment/{expId}/runs, mlflow://registered-model/{name}/versions, mlflow://trace/{traceId}.
Token-efficient by design — extractFields projection on get-run / search-runs / search-traces / get-trace / fat reads, MLFLOW_TOOLS / MLFLOW_DISABLE 8 categories, search-tools meta-tool.
Apps SDK card — compare-runs renders as a side-by-side card on ChatGPT clients (run summary + metric/param tables with diff highlight) via _meta["openai/outputTemplate"]. Claude clients receive the same JSON content.
stdio + Streamable HTTP — defaults to stdio. Set MCP_TRANSPORT=http for ChatGPT Apps SDK or remote clients (Bearer auth via MCP_HTTP_TOKEN).

Related MCP server: MLflow MCP Server

Try this — 5 prompts

Connect the server to Claude Desktop or Claude Code, then paste any of these:

Best run — "In the customer-churn-v3 experiment, find the run with the highest val_accuracy. Show its hyperparameters and metric history."
Failure mode clustering — "Find traces with status=ERROR from the last 24h in experiment 12. Group the failures by exception type and surface the 3 most common."
Run comparison — "Compare the top 5 runs of experiment 12 by validation_loss. Show differing hyperparameters in a table."
Model promotion — "Get the latest version of recommendation_v2 registered model with the champion alias. Show its training metrics + lineage to the source run."
Trace deep-dive — "Pull trace tr-abc123. Highlight slow spans and any failed feedback annotations." (Add list-trace-attachments on Databricks workspaces.)

When to use this vs alternatives

	Official `mlflow[mcp]`	kkruglik/mlflow-mcp	`@us-all/mlflow-mcp` (this)
Tool count	~9 (trace-only)	~25	78
MLflow 3 LoggedModel	❌	✅	✅
Trace attachments	❌	❌	✅ Databricks only
Prompt-optimization-jobs	❌	❌	✅
Webhooks	❌	❌	✅
Aggregation tools	❌	❌	✅ `summarize-experiment`, `summarize-run`
MCP Prompts	❌	✅	✅
MCP Resources	❌	❌	✅ 6 URIs
Auth	Databricks SDK	Bearer / basic	Bearer / basic
Transport	stdio	stdio	stdio

The official mlflow[mcp] is bundled inside MLflow itself and intentionally trace-narrow. Use it for quick managed-MLflow trace inspection. Use this server for end-to-end coverage, especially MLflow 3 entities, prompt-optimization workflows, and aggregation-driven AI debugging.

Install

Claude Desktop

{
  "mcpServers": {
    "mlflow": {
      "command": "npx",
      "args": ["-y", "@us-all/mlflow-mcp"],
      "env": {
        "MLFLOW_TRACKING_URI": "http://localhost:5000"
      }
    }
  }
}

Claude Code

claude mcp add mlflow -s user \
  -e MLFLOW_TRACKING_URI=http://localhost:5000 \
  -- npx -y @us-all/mlflow-mcp

Docker

docker run --rm -i \
  -e MLFLOW_TRACKING_URI=http://your-host:5000 \
  ghcr.io/us-all/mlflow-mcp-server

Build from source

git clone https://github.com/us-all/mlflow-mcp-server.git
cd mlflow-mcp-server && pnpm install && pnpm build
node dist/index.js

Configuration

Variable	Required	Default	Description
`MLFLOW_TRACKING_URI`	✅	—	MLflow tracking URL (`http://localhost:5000`, Databricks workspace URL, etc.)
`MLFLOW_TRACKING_TOKEN`	❌	—	Bearer token. Use for Databricks PAT (`dapi…`)
`MLFLOW_TRACKING_USERNAME`	❌	—	Basic-auth username (alternative to token)
`MLFLOW_TRACKING_PASSWORD`	❌	—	Basic-auth password
`MLFLOW_EXPERIMENT_ID`	❌	—	Default experiment ID for tools that accept it implicitly
`MLFLOW_ALLOW_WRITE`	❌	`false`	Set `true` to enable mutations (create/update/delete)
`MLFLOW_TOOLS`	❌	—	Comma-sep allowlist of categories. Biggest token saver.
`MLFLOW_DISABLE`	❌	—	Comma-sep denylist. Ignored when `MLFLOW_TOOLS` is set.
`MCP_TRANSPORT`	❌	`stdio`	`http` to enable Streamable HTTP transport
`MCP_HTTP_TOKEN`	conditional	—	Bearer token. Required when `MCP_TRANSPORT=http`
`MCP_HTTP_PORT`	❌	`3000`	HTTP listen port
`MCP_HTTP_HOST`	❌	`127.0.0.1`	HTTP bind host (DNS rebinding protection auto-enabled for localhost)
`MCP_HTTP_SKIP_AUTH`	❌	`false`	Skip Bearer auth — e.g. behind a reverse proxy that handles it

Categories (8): experiments, runs, registry, logged-models, traces, assessments, webhooks, prompts.

When MCP_TRANSPORT=http: POST /mcp (Bearer-auth JSON-RPC) + GET /health (public liveness).

Databricks managed MLflow

For Databricks-hosted MLflow:

MLFLOW_TRACKING_URI=https://<workspace>.cloud.databricks.com
MLFLOW_TRACKING_TOKEN=dapi...   # PAT or service-principal token

The MLflow REST API path (/api/2.0/mlflow/...) is identical between OSS and Databricks. Bearer auth handles both PAT and service-principal flows.

Token efficiency

Scenario	Tools	Schema tokens	vs default
default (all categories)	78	9,200	—
typical (`MLFLOW_TOOLS=experiments,runs,registry,traces`)	54	5,900	−36%
narrow (`MLFLOW_TOOLS=experiments,runs`)	27	3,200	−66%

Plus extractFields on get-run / search-runs / search-traces / get-trace / summarize-experiment — caller can scope response fields per call.

Read-only mode

By default, all writes are blocked. The following require MLFLOW_ALLOW_WRITE=true:

create-experiment, update-experiment, delete-experiment, restore-experiment, set-experiment-tag, delete-experiment-tag, create-run, update-run, delete-run, restore-run, log-metric, log-param, log-batch, log-inputs, set-run-tag, delete-run-tag, create-registered-model, rename-registered-model, update-registered-model, delete-registered-model, plus all model-version, logged-model, trace, assessment, webhook, and prompt-optimization writes.

Limitations & gotchas

search-traces.maxResults is clamped to 500. MLflow 3.12+ rejects per-page max_results > 500 with INVALID_PARAMETER_VALUE. For larger result sets, loop on nextPageToken — total trace count is unbounded.
Trace attachments are Databricks-only. list-trace-attachments / get-trace-attachment call routes that OSS MLflow (verified through 3.12.0) returns 404 for. Tool descriptions surface this; calls against OSS return a structured MlflowError.
search-traces.maxResults cap applies per page, not per call — pagination still gets you the full set.
Bearer / Basic auth code paths are not yet validated against live Databricks (see open roadmap item). Works against OSS MLflow 3.12 (Bearer optional).

MCP Prompts (4)

Workflow templates available via MCP prompts/list:

debug-failed-traces — find failed traces, group failure modes
promote-best-run — find best run, register, set champion alias
compare-top-runs — top-N comparison by metric
annotate-trace-quality — guided feedback annotation loop

MCP Resources

URI-based read-only access:

mlflow://run/{runId}, mlflow://experiment/{expId}, mlflow://experiment-by-name/{name}, mlflow://registered-model/{name}, mlflow://model-version/{name}/{version}, mlflow://trace/{traceId}, mlflow://run/{runId}/artifacts, mlflow://experiment/{expId}/runs, mlflow://registered-model/{name}/versions.

Tools (82)

8 categories. Use search-tools to discover at runtime; full list collapsed below.

get-run, search-runs, search-traces, get-trace, and summarize-experiment accept extractFields for response slicing.

Experiments (9)

create-experiment, search-experiments, get-experiment, get-experiment-by-name, update-experiment, delete-experiment, restore-experiment, set-experiment-tag, delete-experiment-tag

Runs (18)

create-run, get-run, search-runs, update-run, delete-run, restore-run, log-metric, log-param, log-batch, log-inputs, get-metric-history, set-run-tag, delete-run-tag, list-artifacts, get-best-run, compare-runs, search-runs-by-tags, summarize-run (aggregation)

Registered Models (12)

create-registered-model, get-registered-model, search-registered-models, rename-registered-model, update-registered-model, delete-registered-model, get-latest-model-versions, set-registered-model-tag, delete-registered-model-tag, set-registered-model-alias, delete-registered-model-alias, get-model-version-by-alias

Model Versions (9)

create-model-version, get-model-version, search-model-versions, update-model-version, delete-model-version, transition-model-version-stage, get-model-version-download-uri, set-model-version-tag, delete-model-version-tag

Logged Models — MLflow 3 (8)

create-logged-model, search-logged-models, get-logged-model, finalize-logged-model, delete-logged-model, set-logged-model-tags, delete-logged-model-tag, log-logged-model-params

Traces (8)

search-traces, get-trace, get-trace-info, delete-traces, set-trace-tag, delete-trace-tag, list-trace-attachments, get-trace-attachment

Assessments (5)

log-feedback, log-expectation, get-assessment, update-assessment, delete-assessment

Webhooks (6)

create-webhook, list-webhooks, get-webhook, update-webhook, delete-webhook, test-webhook

Prompt Optimization (5)

create-prompt-optimization-job, get-prompt-optimization-job, search-prompt-optimization-jobs, cancel-prompt-optimization-job, delete-prompt-optimization-job

Aggregations

summarize-experiment, summarize-run — fold 3–5 round-trips into one structured response with caveats array.

Local validation with docker compose

# 1. start MLflow (UI at http://localhost:5050)
docker compose up -d mlflow

# 2. seed demo experiment, runs, registered model, traces
docker compose run --rm seed

# 3a. probe the MCP server locally against the compose'd MLflow
MLFLOW_TRACKING_URI=http://localhost:5050 \
  MLFLOW_EXPERIMENT_ID=1 \
  MLFLOW_ALLOW_WRITE=true \
  node dist/index.js

# 3b. or run inside compose (stdio)
docker compose run --rm mcp

# tear down
docker compose down -v

./dev/seed.py is idempotent — skips if demo experiment already has runs.

Architecture

Claude → MCP stdio → src/index.ts → src/tools/*.ts → MlflowClient (fetch) → MLflow REST API

Built on @us-all/mcp-toolkit:

extractFields — token-efficient response projections
aggregate(fetchers, caveats) — fan-out helper for summarize-experiment
createWrapToolHandler — Bearer/basic credential redaction + MlflowError extraction
search-tools meta-tool

Targets MLflow 3.5.1+ (uses v3 traces/assessments REST). Dev compose pinned to MLflow 3.12.0 (multimodal trace attachments + paginated trace search).

Tech stack

Node.js 22+ • TypeScript strict ESM • pnpm • @modelcontextprotocol/sdk • zod • dotenv • vitest.

References

MLflow MCP overview: https://mlflow.org/docs/latest/genai/mcp/
MLflow REST API: https://mlflow.org/docs/latest/api_reference/rest-api.html

License

MIT

Available Tools

82 tools

cancel-prompt-optimization-jobB

Destructive

Cancel a running prompt optimization job

Inputs: jobId (required).

compare-runsA

Read-only

Side-by-side metric/param comparison across multiple runs. Renders an Apps SDK card on ChatGPT clients (Claude clients receive the same JSON text).

Inputs: runIds (required), metricKeys, paramKeys.

create-experimentB

Create a new MLflow experiment

Inputs: name (required), artifactLocation, tags.

create-logged-modelC

Create a new MLflow 3 LoggedModel entity in an experiment

Inputs: experimentId, name, modelType, sourceRunId, +2 more.

create-model-versionC

Create a new model version

Inputs: name (required), source (required), runId, tags, +2 more.

create-prompt-optimization-jobB

Create a prompt optimization job to automatically improve a registered prompt

Inputs: sourcePromptUri (required), experimentId, config, tags.

create-registered-modelC

Create a new registered model in the model registry

Inputs: name (required), description, tags.

create-runA

Create a new run in an experiment

Inputs: experimentId, startTime, runName, tags.

create-webhookB

Inputs: name (required), events (required), url (required), description, +3 more.

delete-assessmentB

Destructive

Delete an assessment from a trace

Inputs: traceId (required), assessmentId (required).

delete-experimentA

Destructive

Soft-delete an experiment by ID

Inputs: experimentId (required).

delete-experiment-tagB

Destructive

Delete a tag from an experiment

Inputs: experimentId (required), key (required).

delete-logged-modelA

Destructive

Soft-delete a LoggedModel by ID

Inputs: modelId (required).

delete-logged-model-tagC

Destructive

Delete a tag from a LoggedModel

Inputs: modelId (required), key (required).

delete-model-versionC

Destructive

Delete a specific model version

Inputs: name (required), version (required).

delete-model-version-tagB

Destructive

Delete a tag from a model version

Inputs: name (required), version (required), key (required).

delete-prompt-optimization-jobC

Destructive

Delete a prompt optimization job

Inputs: jobId (required).

delete-registered-modelA

Destructive

Delete a registered model and all its versions

Inputs: name (required).

delete-registered-model-aliasC

Destructive

Delete an alias from a registered model

Inputs: name (required), alias (required).

delete-registered-model-tagC

Destructive

Delete a tag from a registered model

Inputs: name (required), key (required).

delete-runA

Destructive

Soft-delete a run by ID

Inputs: runId (required).

delete-run-tagC

Destructive

Delete a tag from a run

Inputs: runId (required), key (required).

delete-tracesA

Destructive

Delete traces by ID or older than a timestamp

Inputs: experimentId, traceIds, maxTimestampMillis, maxTraces.

delete-trace-tagB

Destructive

Remove a tag from a trace

Inputs: traceId (required), key (required).

delete-webhookA

Destructive

Delete a webhook by ID

Inputs: id (required).

finalize-logged-modelA

Set a terminal status (READY/FAILED/...) on a LoggedModel

Inputs: modelId (required), status (required).

get-assessmentB

Read-only

Get an assessment by trace ID and assessment ID

Inputs: traceId (required), assessmentId (required).

get-best-runA

Read-only

Find the run with the best (max/min) value of a metric in an experiment

Inputs: metric (required), experimentId, ascending, filter.

get-experimentA

Read-only

Get experiment details by ID

Inputs: experimentId, extractFields.

get-experiment-by-nameB

Read-only

Get experiment details by name

Inputs: experimentName (required).

get-latest-model-versionsB

Read-only

Get the latest model versions per stage

Inputs: name (required), stages.

get-logged-modelA

Read-only

Get a LoggedModel by ID

Inputs: modelId (required), allowDeleted.

get-metric-historyB

Read-only

Get full history of a metric for a run

Inputs: runId (required), metricKey (required), pageToken, maxResults.

get-model-versionB

Read-only

Get a model version by name and version

Inputs: name (required), version (required).

get-model-version-by-aliasC

Read-only

Get the model version pointed to by an alias

Inputs: name (required), alias (required).

get-model-version-download-uriC

Read-only

Get the artifact download URI for a model version

Inputs: name (required), version (required).

get-prompt-optimization-jobB

Read-only

Get a prompt optimization job by ID

Inputs: jobId (required).

get-registered-modelB

Read-only

Get a registered model by name

Inputs: name (required).

get-runA

Read-only

Get run details by ID

Inputs: runId (required), extractFields.

get-traceB

Read-only

Retrieve detailed trace information by trace ID

Inputs: traceId (required), extractFields.

get-trace-attachmentA

Read-only

Get a specific attachment on a trace by ID (Databricks MLflow only — OSS servers return 404)

Inputs: traceId (required), attachmentId (required).

get-trace-infoA

Read-only

Retrieve trace metadata only (no spans)

Inputs: traceId (required).

get-webhookA

Read-only

Get webhook details by ID

Inputs: id (required).

list-artifactsB

Read-only

List artifacts under a run's artifact directory

Inputs: runId (required), path, pageToken.

list-trace-attachmentsA

Read-only

List attachments on a trace (Databricks MLflow only — OSS servers return 404)

Inputs: traceId (required).

list-webhooksA

Read-only

List webhooks (optionally filtered by model name)

Inputs: modelName, maxResults, pageToken.

log-batchB

Log a batch of metrics, params, and tags to a run

Inputs: runId (required), metrics, params, tags.

log-expectationC

Log a ground-truth expectation on a trace

Inputs: traceId (required), name (required), value (required), rationale, +2 more.

log-feedbackB

Log evaluation feedback (score or judgment) on a trace

Inputs: traceId (required), name (required), value (required), rationale, +2 more.

log-inputsC

Log dataset inputs to a run

Inputs: runId (required), datasets (required).

log-logged-model-paramsC

Log parameters on a LoggedModel

Inputs: modelId (required), params (required).

log-metricB

Log a single metric value to a run

Inputs: runId (required), key (required), value (required), timestamp, +1 more.

log-paramB

Log a single parameter to a run

Inputs: runId (required), key (required), value (required).

rename-registered-modelA

Rename a registered model

Inputs: name (required), newName (required).

restore-experimentA

Restore a deleted experiment

Inputs: experimentId (required).

restore-runA

Restore a deleted run

Inputs: runId (required).

search-experimentsC

Read-only

Search experiments with filter and pagination

Inputs: filter, maxResults, orderBy, pageToken, +2 more.

search-logged-modelsA

Read-only

Search LoggedModels by experiment with filter and pagination

Inputs: experimentIds, filter, maxResults, orderBy, +1 more.

search-model-versionsB

Read-only

Search model versions with filter and pagination

Inputs: filter, maxResults, orderBy, pageToken.

search-prompt-optimization-jobsB

Read-only

Search prompt optimization jobs in an experiment

Inputs: experimentId, filter, maxResults, pageToken.

search-registered-modelsC

Read-only

Search registered models

Inputs: filter, maxResults, orderBy, pageToken.

search-runsA

Read-only

Search runs with filter expression and pagination

Inputs: experimentIds, filter, runViewType, maxResults, +3 more.

search-runs-by-tagsB

Read-only

Find runs whose tags match all of the given key/value pairs

Inputs: tags (required), experimentIds, maxResults, orderBy.

search-toolsA

Read-only

Discover available tools by natural language query. Returns matching tool names + descriptions across all categories. Use this first to navigate the 77+ tool surface efficiently.

Inputs: query (required), category, limit.

search-tracesB

Read-only

Search and filter traces in experiments

Inputs: experimentIds, filter, maxResults, orderBy, +2 more.

set-experiment-tagC

Set a tag on an experiment

Inputs: experimentId (required), key (required), value (required).

set-logged-model-tagsA

Set or upsert tags on a LoggedModel

Inputs: modelId (required), tags (required).

set-model-version-tagC

Set a tag on a model version

Inputs: name (required), version (required), key (required), value (required).

set-registered-model-aliasB

Set an alias on a registered model version

Inputs: name (required), alias (required), version (required).

set-registered-model-tagC

Set a tag on a registered model

Inputs: name (required), key (required), value (required).

set-run-tagC

Set a tag on a run

Inputs: runId (required), key (required), value (required).

set-trace-tagA

Add a custom key-value tag to a trace

Inputs: traceId (required), key (required), value (required).

summarize-experimentA

Read-only

Aggregated experiment view: experiment overview + topN runs (sorted by metric or start_time) + metric stats (min/max/mean across topN) in a single call. Replaces 3-5 round-trips of get-experiment + search-runs + get-best-run.

Inputs: experimentId (required), topN, metric, ascending, +1 more.

summarize-runA

Read-only

Aggregated run view: run info + (optional) metric history + (optional) artifacts list in a single call. Replaces 3-4 round-trips of get-run + get-metric-history (per metric) + list-artifacts.

Inputs: runId (required), includeMetricHistory, metricKeys, includeArtifacts, +2 more.

test-webhookA

Send a test event to a webhook to verify configuration

Inputs: id (required).

transition-model-version-stageC

Transition a model version to a new stage

Inputs: name (required), version (required), stage (required), archiveExistingVersions.

update-assessmentC

Update an existing assessment

Inputs: traceId (required), assessmentId (required), value, rationale, +1 more.

update-experimentB

Rename an experiment

Inputs: experimentId (required), newName (required).

update-model-versionC

Update a model version's description

Inputs: name (required), version (required), description.

update-registered-modelB

Update a registered model's description

Inputs: name (required), description.

update-runC

Update run status, end time, or name

Inputs: runId (required), status, endTime, runName.

update-webhookB

Update an existing webhook (events, URL, status, secret)

Inputs: id (required), events, url, description, +2 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

Claude Code, inside the Godot 4 editorFlockbay puts an AI agent in the real Godot 4 editor: it writes GDScript, runs your game and fixes what breaks. Free, or bring your own Claude Code or Codex.flockbay.comAd

TDQS

B3.2/5.0

Scored 2026-06-19 17:05 across 82 tools

Disambiguation5/5

Every tool targets a distinct resource-action pair (e.g., create-experiment vs create-run vs create-logged-model). Even similar functions like search-runs and search-runs-by-tags are clearly differentiated by the 'by-tags' suffix. There is no ambiguity between any two tools.

Naming Consistency5/5

All tool names follow a consistent verb-noun pattern using lowercase and hyphens (e.g., delete-run-tag, set-experiment-tag, search-registered-models). The pattern is uniform across all 82 tools, with no mixing of conventions or unexpected formats.

Tool Count2/5

With 82 tools, the count far exceeds the recommended 3-15 range for a typical server. While MLflow is a broad platform, the fine-grained separation into individual operations (e.g., separate log-metric, log-param, log-batch) results in an overwhelming surface that could hinder agent efficiency, despite the presence of a search-tools helper.

Completeness5/5

The tool set covers the entire MLflow lifecycle: experiments, runs, model registry (including aliases and versions), logged models, prompt optimization, traces with assessments, webhooks, and compound aggregate tools like summarize-experiment. There are no obvious dead ends or missing CRUD operations for the domain.

Maintenance

ActivityInactive

ResponsivenessOutdated data

Resources

Related MCP Connectors

Ai Model ExperimentsA
io.github.pipeworx-io
AI Model Experiments MCP ('Model Lab') — run the same prompts across many AI models simultaneously…
AI & Machine Learning Research & Data Developer Tools
521
https://gateway.pipeworx.io/ai-model-experiments/mcp
ПромтЛаб — библиотека AI-промптов
ru.promtlabs
Self-hosted AI prompt library: prompts, collections, tags, teams, chains. 29 MCP tools for agents.
AI & Machine Learning Knowledge & Memory Developer Tools
2
https://promtlabs.ru/mcp
Bay RunA
io.github.barneywohl
Free OpenAI-compatible inference with signed provenance receipts and 3 focused MCP tools.
AI & Machine Learning RAG Systems
31
https://run.huggingbay.xyz/mcp/
anotherai
dev.anotherai
MCP server for building and testing AI agents with multi-model experimentation and insights.
Agent Orchestration Autonomous Agents Testing & QA Tools
2
https://api.anotherai.dev/mcp

Related MCP Servers

trackio-mcp
RAG Systems Monitoring Developer Tools
fcakyon
A
license
Not graded
quality
D
maintenance
An MCP server that enables AI agents to observe and interact with trackio experiment tracking, providing tools for managing ML experiments through natural language.
Updated a year ago (2025-08-03 13:39 UTC)
3
MIT
MLflow MCP Server
Research & Data Developer Tools Monitoring
yesid-lopez
A
license
C
quality
D
maintenance
Enables AI assistants to interact with MLflow experiments, runs, and registered models. Supports browsing experiments, retrieving run details with metrics and parameters, and querying the model registry through natural language.
Updated 8 months ago (2026-03-11 00:25 UTC)
7
MIT
FleetQ
Agent Orchestration Autonomous Agents Developer Tools
escapeboy
A
license
A
quality
B
maintenance
AI Agent Mission Control — 200+ MCP tools across 31 domains. Manage agents, experiments, workflows, crews, skills, tools, credentials, approvals, signals, budgets, marketplace, knowledge bases, chatbots, and more. Self-hosted, open-source (AGPL-3.0). Supports stdio + Streamable HTTP/SSE with OAuth 2.0 auth.
Updated a day ago (2026-11-12 01:38 UTC)
34
81
AGPL 3.0
MLflow MCP Server
AI & Machine Learning Developer Tools
kkruglik
A
license
A
quality
C
maintenance
A Model Context Protocol server that enables LLMs to interact with MLflow tracking servers, allowing users to query experiments, analyze runs, compare metrics, manage the model registry, and promote models through natural language.
Updated 3 months ago (2026-08-08 13:03 UTC)
40
296 PyPI
15
MIT

MLflow MCP Server

What it does that others don't

Try this — 5 prompts

When to use this vs alternatives

Install

Claude Desktop

Claude Code

Docker

Build from source

Configuration

Databricks managed MLflow

Token efficiency

Read-only mode

Limitations & gotchas

MCP Prompts (4)

MCP Resources

Tools (82)

Experiments (9)

Runs (18)

Registered Models (12)

Model Versions (9)

Logged Models — MLflow 3 (8)

Traces (8)

Assessments (5)

Webhooks (6)

Prompt Optimization (5)

Aggregations

Meta

Local validation with docker compose

Architecture

Tech stack

References

License

Available Tools

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

trackio-mcp

MLflow MCP Server

FleetQ

MLflow MCP Server