Skip to content

Connectors · AI

Langfuse MCP server for AI agents.

actionlake is an MCP gateway: one remote MCP server that connects Claude and other MCP clients to Langfuse and the rest of your systems. It gives your agent 112 Langfuse actions (50 reads, 40 writes and 22 that wait for a confirmation) and holds the Langfuse credential itself, so the agent never does. It paces Langfuse's limits and records every call.

100,000 free calls a month. No card, no email, no password: press Start a free sandbox on the next screen and keep the key it shows you.

claude mcp add --transport http actionlake https://connect.actionlake.com/mcp
the first call
do {  action: "langfuse.scim_users.list",  intent: "check which Langfuse account is connected"}# List users in the organization (requires organization-scoped API key) · read · 1 upstream call
  • 112actionsin one signed file
  • 50readlook, never change
  • 40writecreate and update
  • 22wait for a yesdestructive, confirmed first

Set up

Connect Langfuse in three steps.

  1. 01

    Add actionlake to Claude.

    Run the command once in Claude Code; its consent page can create a free sandbox: 100,000 calls a month, no card. In the Claude app, add a custom connector with the URL below and paste your key on the consent page. Creating a new sandbox from the Claude app is not yet tested by us.

    https://connect.actionlake.com/mcp
  2. 02

    Connect Langfuse.

    Username and secret. No consent screen: your agent can connect it for you, but the key then passes through your chat. To keep it out of the model's context, paste it in the console instead. actionlake stores what you paste write-only, sealed under your tenant's key.

    • public_keyLangfuse public key (pk-lf-...): in Langfuse open Project Settings > API Keys > Create new API key (or Organization…
    • secret_keyLangfuse secret key (sk-lf-...), shown once when the API key is created
    • hostWhere your Langfuse lives: your host
    account { op: "connections.requirements", service: "langfuse" }
    # free: exactly what to paste, and what is found for you
  3. 03

    Make the first call. Check before a write.

    Ask in plain words; Claude picks the action. Before anything changes, check validates the call and prices it, for free.

    do    { action: "langfuse.scim_users.list" }# list users in the organization (requires organization-scoped api key) · read · 1 upstream callcheck { action: "langfuse.scim_users.create", input: { … } }# free: inputs, permissions, cost and confirmations, nothing sent

Langfuse at a glance

Sign-in
Username and secret. No consent screen: your agent can connect it for you, but the key then passes through your chat. To keep it out of the model's context, paste it in the console instead.
You provide
public_keysecret_key
Setup fields
host
Limits we pace to
1,000 calls a minute, per connection
API covered
Langfuse Public API (OpenAPI 3.0.1)cloud.langfuse.com
Built from
Langfuse's documentation, read 23 October 2026cloud.langfuse.comapi.reference.langfuse.com
Service file
Version 1.0.0, signed (Ed25519)
Cached reads
44 of 50 reads, for up to 5 minutes; a write clears what it affects
Shared entities
3, so ask can line Langfuse up with your other systems
Data
Processed and stored in Sydney, Australia. No US region yet

Capabilities

What an AI agent can do in Langfuse.

112 actions in 40 areas. Every one carries its effect, so a key can allow the reads and hold the rest.

  • 50readlook, never change
  • 40writecreate and update
  • 22destructivewait for a confirmation
  1. Annotation queue items

    5

    2 read · 2 write · 1 destructive

    Add an item to an annotation queue · Get items for a specific annotation queue · Get a specific item from an annotation queue · Update an annotation queue item

  2. Dashboard widgets

    5

    2 read · 2 write · 1 destructive

    Create a dashboard widget (a standalone chart definition you place on any dashboard) · List dashboard widgets in the project, ordered by most recently updated first · Get a dashboard widget by id · Update a dashboard widget

  3. Dashboards

    5

    2 read · 2 write · 1 destructive

    List dashboards in the project, ordered by most recently updated first · Create a dashboard · Get a dashboard by id · Update a dashboard's name, description, definition, or filters

  4. Evaluation rules

    5

    2 read · 2 write · 1 destructive

    Create an evaluation rule using stable evaluator identifiers · List evaluation rules in newest-first creation order · Get one evaluation rule, including a legacy trace or dataset rule, by its stable identifier · Update an evaluation rule by its stable identifier

  5. Evaluators

    5

    2 read · 2 write · 1 destructive

    Create an evaluator in the authenticated project · List evaluators in newest-first creation order · Get one evaluator by its stable identifier · Update an evaluator by its stable identifier

  6. Models

    5

    2 read · 2 write · 1 destructive

    Create a model · Get all models · Get a model · Delete a model

  7. Blob storage integrations

    4

    2 read · 1 write · 1 destructive

    Get all blob storage integrations for the organization (requires organization-scoped API key) · Get the sync status of a blob storage integration by integration ID (requires organization-scoped API key) · Delete a blob storage integration by ID (requires organization-scoped API key) · Create or update a blob storage integration for a specific project (requires organization-scoped API key)

  8. Dataset items

    4

    2 read · 1 write · 1 destructive

    Get dataset items · Get a dataset item · Delete a dataset item and all its run items · Create a dataset item

32 more areas · 74 actions
  • Projects 4
  • Prompts 4
  • Scim users 4
  • Score configs 4
  • Skills 4
  • Annotation queues 3
  • Comments 3
  • Dashboard placements 3
  • Datasets 3
  • LLM connections 3
  • Media 3
  • Organization LLM connections 3
  • Organization memberships 3
  • Project API keys 3
  • Project memberships 3
  • Scim 3
  • Scores 3
  • Annotation queue assignments 2
  • Skill versions 2
  • Traces 2
  • Evaluator versions 1
  • Experiment items 1
  • Experiments 1
  • Feedback 1
  • Health 1
  • Metrics 1
  • Observations 1
  • Organization API keys 1
  • Organization projects 1
  • Otel traces 1
  • Prompt versions 1
  • Skill files 1

Top actions

The Langfuse API, shaped for agents.

The ids your agent types, what each does, and the one name it shares with other systems. 21 of Langfuse's 21 action names are shared with at least one other connector, so an agent learns them once.

ActionWhat it doesEffectSame name in
langfuse.scim_users.listList users in the organization (requires organization-scoped API key)List users in the organization (requires organization-scoped API key).readidentity.user.list474 connectors
langfuse.datasets.listGet all datasetsGet all datasets.readanalytics.dataset.list28 connectors
langfuse.scim_users.createCreate a new user in the organization (requires organization-scoped API key)Create a new user in the organization (requires organization-scoped API key).writeidentity.user.create262 connectors
langfuse.datasets.createCreate a datasetCreate a dataset.writeanalytics.dataset.create21 connectors
langfuse.scim_users.deleteRemove a user from the organization (requires organization-scoped API key)Remove a user from the organization (requires organization-scoped API key).destructiveidentity.user.delete253 connectors
langfuse.project_api_keys.createCreate a new API key for a project (requires organization-scoped API key)Create a new API key for a project (requires organization-scoped API key).writeidentity.api_key.create100 connectors
langfuse.project_api_keys.listGet all API keys for a project (requires organization-scoped API key)Get all API keys for a project (requires organization-scoped API key).readidentity.api_key.list96 connectors
langfuse.dashboards.listList dashboards in the project, ordered by most recently updated firstList dashboards in the project, ordered by most recently updated first.readanalytics.dashboard.list22 connectors
langfuse.dashboards.createCreate a dashboardCreate a dashboard.writeanalytics.dashboard.create18 connectors
langfuse.metrics.queryGet metrics from the Langfuse project using a query objectGet metrics from the Langfuse project using a query object.readanalytics.metric.query5 connectors
langfuse.organization_memberships.listGet all memberships for the organization associated with the API key (requires organization-scoped API key)Get all memberships for the organization associated with the API key (requires organization-scoped API key).readthis connector
langfuse.observations.listGet a list of observations with cursor-based pagination and flexible field selectionGet a list of observations with cursor-based pagination and flexible field selection.readmonitoring.trace.list2 connectors

Try it

Prompts to paste into Claude.

Each one names the Langfuse actions it will call, so you know what it reads and what it changes before you send it.

  1. Readread

    Who has access in Langfuse? List every user and their role.

    callslangfuse.scim_users.list

  2. Writewrite

    Create a new user in the organization (requires organization-scoped API key) in Langfuse. Show me exactly what you will send before you save it.

    callslangfuse.scim_users.create

Governance for Langfuse

Destructive actions wait for a yes.

22 Langfuse actions are destructive. These are all of them:

  • Remove a user from the organization (requires organization-scoped API key)langfuse.scim_users.delete
  • Delete an API key for a project (requires organization-scoped API key)langfuse.project_api_keys.delete
  • Delete a dashboardlangfuse.dashboards.delete
  • Delete a modellangfuse.models.delete
  • Delete a score (supports both trace and session scores)langfuse.scores.delete
  • Delete a specific tracelangfuse.traces.delete
  • Delete prompt versionslangfuse.prompts.delete
  • Delete a project by ID (requires organization-scoped API key)langfuse.projects.delete
  • Delete an evaluator and all of its stored versionslangfuse.evaluators.delete
  • Delete multiple traceslangfuse.traces.delete_many
  • Delete a dataset item and all its run itemslangfuse.dataset_items.delete
  • Delete one immutable skill versionlangfuse.skill_versions.delete
  • Delete an LLM connection by idlangfuse.llm_connections.delete
  • Delete an evaluation rulelangfuse.evaluation_rules.delete
  • Delete a dashboard widgetlangfuse.dashboard_widgets.delete
  • Delete a membership from a specific project (requires organization-scoped API key)langfuse.project_memberships.delete
  • Remove a placement from a dashboard grid without deleting the referenced widgetlangfuse.dashboard_placements.delete
  • Remove an item from an annotation queuelangfuse.annotation_queue_items.delete
  • Delete a membership from the organization associated with the API key (requires organization-scoped API key)langfuse.organization_memberships.delete
  • Delete a blob storage integration by ID (requires organization-scoped API key)langfuse.blob_storage_integrations.delete
  • Delete an assignment for a user to an annotation queuelangfuse.annotation_queue_assignments.delete
  • Delete an organization-level LLM connection by idlangfuse.organization_llm_connections.delete

A destructive action runs only with confirm: true. Under the human policy the agent cannot confirm for itself: the run stops until a named person approves, in the console or from a signed link in an email. The cap on a key is refused before the call is sent, never after.

dolangfuse.scim_users.deleteconfirm_policy: human

Remove a user from the organization (requires organization-scoped API key)?

intent: the user asked me to remove a user from the organization (requires organization-scoped API key)

run suspended · waiting for a named person · key langfuse-agent cannot confirm itself
In the console, or from an emailed signed link.
a key for this connector
account {  op: "keys.create", name: "langfuse-agent",  allowed_actions: ["langfuse.*"], cap_calls: 2000,  confirm_policy: "human_for_writes",  approval_rules: [    { match: "langfuse.scim_users.delete", approver: "[email protected]" },    { match: "langfuse.project_api_keys.delete", approver: "[email protected]" },    { match: "langfuse.dashboards.delete", approver: "[email protected]" }  ]}
Illustrative. Scoped to Langfuse only, capped at 2,000 upstream calls, every write confirmed, and the destructive ones approved by a named person.

Cost

What Langfuse costs on actionlake.

Every Langfuse action is one upstream call, the only thing actionlake bills. Searching, check, cost, account and any call Langfuse refused with 401, 403, 429 or a 5xx are free. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000.

  • 1upstream call per actionall 112 Langfuse actions
  • 112actions pageeach page of a long list is one call
  • 44reads can be cacheda cache hit costs nothing
  • 100,000calls a month, freepaid plans from US$49 a month for 250,000

See pricing

Action index

Every Langfuse action.

112 actions by area: the id your agent calls, what it does, and its effect.

Annotation queue items5
  • langfuse.annotation_queue_items.createAdd an item to an annotation queuewrite
  • langfuse.annotation_queue_items.deleteRemove an item from an annotation queuedestructive
  • langfuse.annotation_queue_items.getGet a specific item from an annotation queueread
  • langfuse.annotation_queue_items.listGet items for a specific annotation queueread
  • langfuse.annotation_queue_items.updateUpdate an annotation queue itemwrite
Dashboard widgets5
  • langfuse.dashboard_widgets.createCreate a dashboard widget (a standalone chart definition you place on any dashboard)write
  • langfuse.dashboard_widgets.deleteDelete a dashboard widgetdestructive
  • langfuse.dashboard_widgets.getGet a dashboard widget by idread
  • langfuse.dashboard_widgets.listList dashboard widgets in the project, ordered by most recently updated firstread
  • langfuse.dashboard_widgets.updateUpdate a dashboard widgetwrite
Dashboards5
  • langfuse.dashboards.createCreate a dashboardwrite
  • langfuse.dashboards.deleteDelete a dashboarddestructive
  • langfuse.dashboards.getGet a dashboard by idread
  • langfuse.dashboards.listList dashboards in the project, ordered by most recently updated firstread
  • langfuse.dashboards.updateUpdate a dashboard's name, description, definition, or filterswrite
Evaluation rules5
  • langfuse.evaluation_rules.createCreate an evaluation rule using stable evaluator identifierswrite
  • langfuse.evaluation_rules.deleteDelete an evaluation ruledestructive
  • langfuse.evaluation_rules.getGet one evaluation rule, including a legacy trace or dataset rule, by its stable identifierread
  • langfuse.evaluation_rules.listList evaluation rules in newest-first creation orderread
  • langfuse.evaluation_rules.updateUpdate an evaluation rule by its stable identifierwrite
Evaluators5
  • langfuse.evaluators.createCreate an evaluator in the authenticated projectwrite
  • langfuse.evaluators.deleteDelete an evaluator and all of its stored versionsdestructive
  • langfuse.evaluators.getGet one evaluator by its stable identifierread
  • langfuse.evaluators.listList evaluators in newest-first creation orderread
  • langfuse.evaluators.updateUpdate an evaluator by its stable identifierwrite
Models5
  • langfuse.models.createCreate a modelwrite
  • langfuse.models.deleteDelete a modeldestructive
  • langfuse.models.getGet a modelread
  • langfuse.models.listGet all modelsread
  • langfuse.models.upsertCreate or replace a project-owned model using its idwrite
Blob storage integrations4
  • langfuse.blob_storage_integrations.deleteDelete a blob storage integration by ID (requires organization-scoped API key)destructive
  • langfuse.blob_storage_integrations.get_statusGet the sync status of a blob storage integration by integration ID (requires organization-scoped API key)read
  • langfuse.blob_storage_integrations.listGet all blob storage integrations for the organization (requires organization-scoped API key)read
  • langfuse.blob_storage_integrations.upsertCreate or update a blob storage integration for a specific project (requires organization-scoped API key)write
Dataset items4
  • langfuse.dataset_items.createCreate a dataset itemwrite
  • langfuse.dataset_items.deleteDelete a dataset item and all its run itemsdestructive
  • langfuse.dataset_items.getGet a dataset itemread
  • langfuse.dataset_items.listGet dataset itemsread
Projects4
  • langfuse.projects.createCreate a new project (requires organization-scoped API key)write
  • langfuse.projects.deleteDelete a project by ID (requires organization-scoped API key)destructive
  • langfuse.projects.listGet Project associated with API key (requires project-scoped API key)read
  • langfuse.projects.updateUpdate a project by ID (requires organization-scoped API key)write
Prompts4
  • langfuse.prompts.createCreate a new version for the prompt with the given `name`write
  • langfuse.prompts.deleteDelete prompt versionsdestructive
  • langfuse.prompts.getGet a promptread
  • langfuse.prompts.listGet a list of prompt names with versions and labelsread
Scim users4
  • langfuse.scim_users.createCreate a new user in the organization (requires organization-scoped API key)write
  • langfuse.scim_users.deleteRemove a user from the organization (requires organization-scoped API key)destructive
  • langfuse.scim_users.getGet a specific user by ID (requires organization-scoped API key)read
  • langfuse.scim_users.listList users in the organization (requires organization-scoped API key)read
Score configs4
  • langfuse.score_configs.createCreate a score configuration (config)write
  • langfuse.score_configs.getGet a score configread
  • langfuse.score_configs.listGet all score configsread
  • langfuse.score_configs.updateUpdate a score configwrite
Skills4
  • langfuse.skills.create_versionCreate skill versionwrite
  • langfuse.skills.getResolve a skill's metadata and file manifest by version or labelread
  • langfuse.skills.listList skills with metadata and timestamps from their latest version, shared tags, and the version assigned to production…read
  • langfuse.skills.update_tagsReplace the shared tags across all versions of a skillwrite
Annotation queues3
  • langfuse.annotation_queues.createCreate an annotation queuewrite
  • langfuse.annotation_queues.getGet an annotation queue by IDread
  • langfuse.annotation_queues.listGet all annotation queuesread
Comments3
  • langfuse.comments.createCreate a commentwrite
  • langfuse.comments.getGet a comment by idread
  • langfuse.comments.listGet all commentsread
Dashboard placements3
  • langfuse.dashboard_placements.createAdd a placement to a dashboard grid (see `DashboardPlacement` for grid semantics)write
  • langfuse.dashboard_placements.deleteRemove a placement from a dashboard grid without deleting the referenced widgetdestructive
  • langfuse.dashboard_placements.updateMove or resize a placementwrite
Datasets3
  • langfuse.datasets.createCreate a datasetwrite
  • langfuse.datasets.getGet a datasetread
  • langfuse.datasets.listGet all datasetsread
LLM connections3
  • langfuse.llm_connections.deleteDelete an LLM connection by iddestructive
  • langfuse.llm_connections.listGet all LLM connections in a projectread
  • langfuse.llm_connections.upsertCreate or update an LLM connectionwrite
Media3
  • langfuse.media.create_upload_urlGet a presigned upload URL for a media recordwrite
  • langfuse.media.getGet a media recordread
  • langfuse.media.updatePatch a media recordwrite
Organization LLM connections3
  • langfuse.organization_llm_connections.deleteDelete an organization-level LLM connection by iddestructive
  • langfuse.organization_llm_connections.listGet all organization-level LLM connectionsread
  • langfuse.organization_llm_connections.upsertCreate or update an organization-level LLM connectionwrite
Organization memberships3
  • langfuse.organization_memberships.deleteDelete a membership from the organization associated with the API key (requires organization-scoped API key)destructive
  • langfuse.organization_memberships.listGet all memberships for the organization associated with the API key (requires organization-scoped API key)read
  • langfuse.organization_memberships.upsertCreate or update a membership for the organization associated with the API key (requires organization-scoped API key)write
Project API keys3
  • langfuse.project_api_keys.createCreate a new API key for a project (requires organization-scoped API key)write
  • langfuse.project_api_keys.deleteDelete an API key for a project (requires organization-scoped API key)destructive
  • langfuse.project_api_keys.listGet all API keys for a project (requires organization-scoped API key)read
Project memberships3
  • langfuse.project_memberships.deleteDelete a membership from a specific project (requires organization-scoped API key)destructive
  • langfuse.project_memberships.listGet all memberships for a specific project (requires organization-scoped API key)read
  • langfuse.project_memberships.upsertCreate or update a membership for a specific project (requires organization-scoped API key)write
Scim3
  • langfuse.scim.get_resource_typesGet SCIM Resource Types (requires organization-scoped API key)read
  • langfuse.scim.get_schemasGet SCIM Schemas (requires organization-scoped API key)read
  • langfuse.scim.get_service_provider_configGet SCIM Service Provider Configuration (requires organization-scoped API key)read
Scores3
  • langfuse.scores.createCreate scores asynchronously from a list of scoreswrite
  • langfuse.scores.deleteDelete a score (supports both trace and session scores)destructive
  • langfuse.scores.listGet a list of scores with a polymorphic `value` field (v3)read
Annotation queue assignments2
  • langfuse.annotation_queue_assignments.createCreate an assignment for a user to an annotation queuewrite
  • langfuse.annotation_queue_assignments.deleteDelete an assignment for a user to an annotation queuedestructive
Skill versions2
  • langfuse.skill_versions.deleteDelete one immutable skill versiondestructive
  • langfuse.skill_versions.set_labelsReplace the labels on a skill version and atomically move them from other versionswrite
Traces2
  • langfuse.traces.deleteDelete a specific tracedestructive
  • langfuse.traces.delete_manyDelete multiple tracesdestructive
Evaluator versions1
  • langfuse.evaluator_versions.listList an evaluator's version history in newest-first orderread
Experiment items1
  • langfuse.experiment_items.listList experiment items with cursor-based paginationread
Experiments1
  • langfuse.experiments.listList experiments with cursor-based paginationread
Feedback1
  • langfuse.feedback.submitSubmit explicit user-approved feedback about Langfuse skills, MCP tools, CLI, docs, or public APIwrite
Health1
  • langfuse.health.getCheck health of API and databaseread
Metrics1
  • langfuse.metrics.queryGet metrics from the Langfuse project using a query objectread
Observations1
  • langfuse.observations.listGet a list of observations with cursor-based pagination and flexible field selectionread
Organization API keys1
  • langfuse.organization_api_keys.listGet all API keys for the organization associated with the API key (requires organization-scoped API key)read
Organization projects1
  • langfuse.organization_projects.listGet all projects for the organization associated with the API key (requires organization-scoped API key)read
Otel traces1
  • langfuse.otel_traces.export**OpenTelemetry Traces Ingestion Endpoint** This endpoint implements the OTLP/HTTP specification for trace ingestion, p…write
Prompt versions1
  • langfuse.prompt_versions.update_labelsUpdate labels for a specific prompt versionwrite
Skill files1
  • langfuse.skill_files.get_contentsRead a batch of text contents by canonical base64-encoded SHA-256 hashesread

Questions

The Langfuse MCP server, answered.

Is there a Langfuse MCP server for Claude?

Yes. actionlake is an MCP gateway: a remote MCP server that gives Claude, or another MCP client, 112 Langfuse actions across annotation queue items, dashboard widgets, dashboards and evaluation rules, through the same eight tools it uses for every other connector. actionlake is independent; it is not made or endorsed by Langfuse.

What do I need to connect Langfuse?

Langfuse public key (pk-lf-...): in Langfuse open Project Settings > API Keys > Create new API key (or Organization… and Langfuse secret key (sk-lf-...), shown once when the API key is created. Plus your host. The free call account { op: "connections.requirements", service: "langfuse" } lists exactly what to paste.

Can an AI agent delete or change things in Langfuse?

Only if its key allows it, and only with an explicit confirmation. 22 Langfuse actions are destructive, including remove a user from the organization (requires organization-scoped API key), delete an API key for a project (requires organization-scoped API key) and delete a dashboard. Each runs only with confirm: true. Under the human confirm policy, or an approval rule on the key, the run waits until a named person approves it.

Will an AI agent use up my Langfuse API limit?

actionlake paces every call to the limits in the Langfuse service file: 1,000 calls a minute per connection. Repeated reads are answered from cache (44 Langfuse actions can be), which never reaches Langfuse.

What does a Langfuse action cost?

One upstream call, the only unit actionlake bills. A list that pages bills one call per page (112 Langfuse actions page). Cache hits, searches, dry runs and calls Langfuse refused with 401, 403, 429 or a 5xx are free. The Free plan includes 100,000 calls a month with no card. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000.

Which Langfuse API does it cover?

The Langfuse Public API (OpenAPI 3.0.1) at cloud.langfuse.com by default, or the base URL set when you connect. It is built from Langfuse's documentation as read on 23 October 2026. 112 actions in 40 areas, each one in the signed service file (version 1.0.0).

Does actionlake keep my Langfuse data, and where?

actionlake processes and stores data in Sydney, Australia; there is no US region yet. Reads may be cached for up to 5 minutes, and a write clears the reads it affects. Each run's request and response is kept, sealed, for your plan's trace window: 1 day on Free, 3 days on Starter, 7 days on Pro, 14 days on Team, 30 days on Business and 30 days on Enterprise. Your Langfuse credential is sealed under your tenant's key, never logged and never returned. Your traffic is not used to train models.

Which AI clients can use it?

Claude Code, tested by us end to end, and the Claude app as a custom connector, tested by us with an existing key. Other MCP clients that support remote servers over Streamable HTTP (MCP revisions 2025-06-18 and 2025-11-25) and OAuth dynamic client registration, or a bearer header, should connect the same way; we have not tested them. Or plain REST, under the same action names.

Give your agent Langfuse today.

Start free - 100,000 calls/month, no card. No signup. Hard caps on every key. Cached answers are free.