Connectors · Data & integration
Apache Airflow MCP server for AI agents.
actionlake is an MCP gateway: one remote MCP server that connects Claude and other MCP clients to Apache Airflow and the rest of your systems. It gives your agent 125 Apache Airflow actions (69 reads, 37 writes and 19 that wait for a confirmation) and holds the Apache Airflow credential itself, so the agent never does. It paces Apache Airflow'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/mcpdo { action: "apache_airflow.dags.list", intent: "check which Apache Airflow account is connected"}# Get Dags · read · 1 upstream call- 125actionsin one signed file
- 69readlook, never change
- 37writecreate and update
- 19wait for a yesdestructive, confirmed first
Set up
Connect Apache Airflow in three steps.
- 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 - 02
Connect Apache Airflow.
Session login. actionlake opens the session and renews it when it expires. actionlake stores what you paste write-only, sealed under your tenant's key.
usernameUsername of the Airflow account actionlake signs in aspasswordPassword of that Airflow account. actionlake exchanges it at /auth/token for a JWThostWhere your Apache Airflow lives: your host
account { op: "connections.requirements", service: "apache_airflow" } # free: exactly what to paste, and what is found for you - 03
Make the first call. Check before a write.
Ask in plain words; Claude picks the action. Before anything changes,
checkvalidates the call and prices it, for free.do { action: "apache_airflow.dags.list" }# get dags · read · 1 upstream callcheck { action: "apache_airflow.connection.create", input: { … } }# free: inputs, permissions, cost and confirmations, nothing sent
Apache Airflow at a glance
- Sign-in
- Session login. actionlake opens the session and renews it when it expires.
- You provide
usernamepassword- Setup fields
host- Limits we pace to
- 10 calls a second, per connection
- API covered
- Airflow 3.3.2 stable REST API v2 (OpenAPI 3.1), at your Apache Airflow instance's API base URL, which you enter when connecting
- Built from
- Apache Airflow's documentation, read 23 October 2026OpenAPI spec (GitHub)airflow.apache.org
- Service file
- Version 1.0.0, signed (Ed25519)
- Cached reads
- 58 of 69 reads, for up to 5 minutes; a write clears what it affects
- Shared entities
- 2, so ask can line Apache Airflow up with your other systems
- Data
- Processed and stored in Sydney, Australia. No US region yet
Capabilities
What an AI agent can do in Apache Airflow.
125 actions in 70 areas. Every one carries its effect, so a key can allow the reads and hold the rest.
- 69readlook, never change
- 37writecreate and update
- 19destructivewait for a confirmation
Backfill
62 read · 3 write · 1 destructive
Create Backfill · Create Backfill Dry Run · Get Backfill · Pause Backfill
Connection
52 read · 2 write · 1 destructive
Post Connection · Get Connection · Patch Connection · Delete Connection
Dag
51 read · 3 write · 1 destructive
Get Dag · Patch Dag · Delete Dag · Favorite Dag
Dag run
51 read · 3 write · 1 destructive
Get Dag Run · Trigger Dag Run · Patch Dag Run · Clear Dag Run
Dag runs
42 read · 1 write · 1 destructive
Get Dag Runs · Get List Dag Runs Batch · Clear Dag Runs · Bulk Dag Runs
Pool
41 read · 2 write · 1 destructive
Post Pool · Get Pool · Patch Pool · Delete Pool
Task instance
41 read · 2 write · 1 destructive
Get Task Instance · Patch Task Instance · Patch Task Instance Dry Run · Delete Task Instance
Task instances
42 read · 2 write · 0 destructive
Get Task Instances Batch · Get Task Instances · Bulk Task Instances · Post Clear Task Instances
62 more areas · 88 actions
- Task state store entry 4
- Variable 4
- Xcom entry 4
- Asset state store entry 3
- Mapped task instance 3
- Asset 2
- Asset queued events 2
- Asset state store 2
- Connection test 2
- Connections 2
- Dag asset queued event 2
- Dag asset queued events 2
- Dags 2
- Hitl detail 2
- Pools 2
- Task group instances 2
- Task state store 2
- Variables 2
- Asset alias 1
- Asset aliases 1
- Asset event 1
- Asset events 1
- Assets 1
- Backfills 1
- Config 1
- Config option 1
- Dag details 1
- Dag file 1
- Dag run partitions 1
- Dag run upstream asset events 1
- Dag source 1
- Dag stats 1
- Dag tags 1
- Dag version 1
- Dag versions 1
- Dag warnings 1
- Default connections 1
- Event log 1
- Event logs 1
- Health 1
- Hitl detail try 1
- Hitl details 1
- Import error 1
- Import errors 1
- Jobs 1
- Mapped task instance dependencies 1
- Mapped task instance tries 1
- Mapped task instance try 1
- Mapped task instances 1
- Plugin import errors 1
- Plugins 1
- Providers 1
- Task 1
- Task instance dependencies 1
- Task instance external log URL 1
- Task instance extra links 1
- Task instance log 1
- Task instance tries 1
- Task instance try 1
- Tasks 1
- Version 1
- Xcom entries 1
Top actions
The Apache Airflow API, shaped for agents.
The ids your agent types, what each does, and the one name it shares with other systems. 15 of Apache Airflow's 15 action names are shared with at least one other connector, so an agent learns them once.
| Action | What it does | Effect | Same name in |
|---|---|---|---|
apache_airflow.dags.listGet Dags | Get all Dags. | read | automation.workflow.list51 connectors |
apache_airflow.dag_runs.listGet Dag Runs | Get all Dag Runs. | read | automation.workflow_run.list27 connectors |
apache_airflow.connections.listGet Connections | Get all connection entries. | read | automation.connection.list32 connectors |
apache_airflow.connection.createPost Connection | Create connection entry. | write | automation.connection.create30 connectors |
apache_airflow.dag.getGet Dag | Get basic information about a Dag. | read | automation.workflow.get51 connectors |
apache_airflow.dag.updatePatch Dag | Patch the specific Dag. | write | automation.workflow.update42 connectors |
apache_airflow.dag_run.getGet Dag Run | Get Dag Run. | read | automation.workflow_run.get27 connectors |
apache_airflow.dag.deleteDelete Dag | Delete the specific Dag. | destructive | automation.workflow.delete40 connectors |
apache_airflow.connection.updatePatch Connection | Update a connection entry. | write | automation.connection.update24 connectors |
apache_airflow.task_instance_log.getGet Log | Get logs for a specific task instance. | read | automation.workflow_run_log.get3 connectors |
apache_airflow.health.getGet Health | Get Health. | read | platform.health.get18 connectors |
apache_airflow.jobs.listGet Jobs | Get all jobs. | read | this connector |
Try it
Prompts to paste into Claude.
Each one names the Apache Airflow actions it will call, so you know what it reads and what it changes before you send it.
- Readread
Get Dags in Apache Airflow and tell me what stands out.
calls
apache_airflow.dags.list - Writewrite
Post Connection in Apache Airflow. Show me exactly what you will send before you save it.
calls
apache_airflow.connection.create
Governance for Apache Airflow
Destructive actions wait for a yes.
19 Apache Airflow actions are destructive. These are all of them:
- Delete Dag
apache_airflow.dag.delete - Delete Connection
apache_airflow.connection.delete - Bulk Pools
apache_airflow.pools.bulk - Delete Pool
apache_airflow.pool.delete - Bulk Dag Runs
apache_airflow.dag_runs.bulk - Delete Dag Run
apache_airflow.dag_run.delete - Bulk Variables
apache_airflow.variables.bulk - Delete Variable
apache_airflow.variable.delete - Bulk Connections
apache_airflow.connections.bulk - Delete Xcom Entry
apache_airflow.xcom_entry.delete - Delete Task Instance
apache_airflow.task_instance.delete - Clear Task State Store
apache_airflow.task_state_store.clear - Clear Asset State Store
apache_airflow.asset_state_store.clear - Delete Asset Queued Events
apache_airflow.asset_queued_events.delete - Delete Dag Asset Queued Event
apache_airflow.dag_asset_queued_event.delete - Delete Task State Store
apache_airflow.task_state_store_entry.delete - Delete Asset State Store
apache_airflow.asset_state_store_entry.delete - Delete Dag Asset Queued Events
apache_airflow.dag_asset_queued_events.delete - Cancel Backfill
apache_airflow.backfill.cancel
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.
Delete Dag?
intent: the user asked me to delete Dag
account { op: "keys.create", name: "apache-airflow-agent", allowed_actions: ["apache_airflow.*"], cap_calls: 2000, confirm_policy: "human_for_writes", approval_rules: [ { match: "apache_airflow.dag.delete", approver: "[email protected]" }, { match: "apache_airflow.connection.delete", approver: "[email protected]" }, { match: "apache_airflow.pools.bulk", approver: "[email protected]" } ]}Cost
What Apache Airflow costs on actionlake.
Every Apache Airflow action is one upstream call, the only thing actionlake bills. Searching, check, cost, account and any call Apache Airflow 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 125 Apache Airflow actions
- 125actions pageeach page of a long list is one call
- 58reads can be cacheda cache hit costs nothing
- 100,000calls a month, freepaid plans from US$49 a month for 250,000
Action index
Every Apache Airflow action.
125 actions by area: the id your agent calls, what it does, and its effect.
Backfill6
apache_airflow.backfill.cancelCancel Backfilldestructiveapache_airflow.backfill.createCreate Backfillwriteapache_airflow.backfill.dry_runCreate Backfill Dry Runreadapache_airflow.backfill.getGet Backfillreadapache_airflow.backfill.pausePause Backfillwriteapache_airflow.backfill.unpauseUnpause Backfillwrite
Connection5
apache_airflow.connection.createPost Connectionwriteapache_airflow.connection.deleteDelete Connectiondestructiveapache_airflow.connection.getGet Connectionreadapache_airflow.connection.testTest Connectionreadapache_airflow.connection.updatePatch Connectionwrite
Dag5
apache_airflow.dag.deleteDelete Dagdestructiveapache_airflow.dag.favoriteFavorite Dagwriteapache_airflow.dag.getGet Dagreadapache_airflow.dag.unfavoriteUnfavorite Dagwriteapache_airflow.dag.updatePatch Dagwrite
Dag run5
apache_airflow.dag_run.clearClear Dag Runwriteapache_airflow.dag_run.deleteDelete Dag Rundestructiveapache_airflow.dag_run.getGet Dag Runreadapache_airflow.dag_run.triggerTrigger Dag Runwriteapache_airflow.dag_run.updatePatch Dag Runwrite
Dag runs4
apache_airflow.dag_runs.batch_listGet List Dag Runs Batchreadapache_airflow.dag_runs.bulkBulk Dag Runsdestructiveapache_airflow.dag_runs.clearClear Dag Runswriteapache_airflow.dag_runs.listGet Dag Runsread
Pool4
apache_airflow.pool.createPost Poolwriteapache_airflow.pool.deleteDelete Pooldestructiveapache_airflow.pool.getGet Poolreadapache_airflow.pool.updatePatch Poolwrite
Task instance4
apache_airflow.task_instance.deleteDelete Task Instancedestructiveapache_airflow.task_instance.getGet Task Instancereadapache_airflow.task_instance.updatePatch Task Instancewriteapache_airflow.task_instance.update_dry_runPatch Task Instance Dry Runwrite
Task instances4
apache_airflow.task_instances.batch_listGet Task Instances Batchreadapache_airflow.task_instances.bulkBulk Task Instanceswriteapache_airflow.task_instances.clearPost Clear Task Instanceswriteapache_airflow.task_instances.listGet Task Instancesread
Task state store entry4
apache_airflow.task_state_store_entry.deleteDelete Task State Storedestructiveapache_airflow.task_state_store_entry.getGet Task State Storereadapache_airflow.task_state_store_entry.setSet Task State Storewriteapache_airflow.task_state_store_entry.updatePatch Task State Storewrite
Variable4
apache_airflow.variable.createPost Variablewriteapache_airflow.variable.deleteDelete Variabledestructiveapache_airflow.variable.getGet Variablereadapache_airflow.variable.updatePatch Variablewrite
Xcom entry4
apache_airflow.xcom_entry.createCreate Xcom Entrywriteapache_airflow.xcom_entry.deleteDelete Xcom Entrydestructiveapache_airflow.xcom_entry.getGet Xcom Entryreadapache_airflow.xcom_entry.updateUpdate Xcom Entrywrite
Asset state store entry3
apache_airflow.asset_state_store_entry.deleteDelete Asset State Storedestructiveapache_airflow.asset_state_store_entry.getGet Asset State Storereadapache_airflow.asset_state_store_entry.setSet Asset State Storewrite
Mapped task instance3
apache_airflow.mapped_task_instance.getGet Mapped Task Instancereadapache_airflow.mapped_task_instance.updatePatch Task Instancewriteapache_airflow.mapped_task_instance.update_dry_runPatch Task Instance Dry Runwrite
Asset2
apache_airflow.asset.getGet Assetreadapache_airflow.asset.materializeMaterialize Assetwrite
Asset queued events2
apache_airflow.asset_queued_events.deleteDelete Asset Queued Eventsdestructiveapache_airflow.asset_queued_events.listGet Asset Queued Eventsread
Asset state store2
apache_airflow.asset_state_store.clearClear Asset State Storedestructiveapache_airflow.asset_state_store.listList Asset State Storeread
Connection test2
apache_airflow.connection_test.enqueueEnqueue Connection Testwriteapache_airflow.connection_test.getGet Connection Testread
Connections2
apache_airflow.connections.bulkBulk Connectionsdestructiveapache_airflow.connections.listGet Connectionsread
Dag asset queued event2
apache_airflow.dag_asset_queued_event.deleteDelete Dag Asset Queued Eventdestructiveapache_airflow.dag_asset_queued_event.getGet Dag Asset Queued Eventread
Dag asset queued events2
apache_airflow.dag_asset_queued_events.deleteDelete Dag Asset Queued Eventsdestructiveapache_airflow.dag_asset_queued_events.listGet Dag Asset Queued Eventsread
Dags2
apache_airflow.dags.listGet Dagsreadapache_airflow.dags.updatePatch Dagswrite
Hitl detail2
apache_airflow.hitl_detail.getGet Hitl Detailreadapache_airflow.hitl_detail.updateUpdate Hitl Detailwrite
Pools2
apache_airflow.pools.bulkBulk Poolsdestructiveapache_airflow.pools.listGet Poolsread
Task group instances2
apache_airflow.task_group_instances.updatePatch Task Group Instanceswriteapache_airflow.task_group_instances.update_dry_runPatch Task Group Instances Dry Runwrite
Task state store2
apache_airflow.task_state_store.clearClear Task State Storedestructiveapache_airflow.task_state_store.listList Task State Storeread
Variables2
apache_airflow.variables.bulkBulk Variablesdestructiveapache_airflow.variables.listGet Variablesread
Asset alias1
apache_airflow.asset_alias.getGet Asset Aliasread
Asset aliases1
apache_airflow.asset_aliases.listGet Asset Aliasesread
Asset event1
apache_airflow.asset_event.createCreate Asset Eventwrite
Asset events1
apache_airflow.asset_events.listGet Asset Eventsread
Assets1
apache_airflow.assets.listGet Assetsread
Backfills1
apache_airflow.backfills.listList Backfillsread
Config1
apache_airflow.config.getGet Configread
Config option1
apache_airflow.config_option.getGet Config Valueread
Dag details1
apache_airflow.dag_details.getGet Dag Detailsread
Dag file1
apache_airflow.dag_file.reparseReparse Dag Filewrite
Dag run partitions1
apache_airflow.dag_run_partitions.clearClear Dag Run Partitionswrite
Dag run upstream asset events1
apache_airflow.dag_run_upstream_asset_events.listGet Upstream Asset Eventsread
Dag source1
apache_airflow.dag_source.getGet Dag Sourceread
Dag stats1
apache_airflow.dag_stats.listGet Dag Statsread
Dag tags1
apache_airflow.dag_tags.listGet Dag Tagsread
Dag version1
apache_airflow.dag_version.getGet Dag Versionread
Dag versions1
apache_airflow.dag_versions.listGet Dag Versionsread
Dag warnings1
apache_airflow.dag_warnings.listList Dag Warningsread
Default connections1
apache_airflow.default_connections.createCreate Default Connectionswrite
Event log1
apache_airflow.event_log.getGet Event Logread
Event logs1
apache_airflow.event_logs.listGet Event Logsread
Health1
apache_airflow.health.getGet Healthread
Hitl detail try1
apache_airflow.hitl_detail_try.getGet Hitl Detail Try Detailread
Hitl details1
apache_airflow.hitl_details.listGet Hitl Detailsread
Import error1
apache_airflow.import_error.getGet Import Errorread
Import errors1
apache_airflow.import_errors.listGet Import Errorsread
Jobs1
apache_airflow.jobs.listGet Jobsread
Mapped task instance dependencies1
apache_airflow.mapped_task_instance_dependencies.listGet Task Instance Dependenciesread
Mapped task instance tries1
apache_airflow.mapped_task_instance_tries.listGet Mapped Task Instance Triesread
Mapped task instance try1
apache_airflow.mapped_task_instance_try.getGet Mapped Task Instance Try Detailsread
Mapped task instances1
apache_airflow.mapped_task_instances.listGet Mapped Task Instancesread
Plugin import errors1
apache_airflow.plugin_import_errors.listImport Errorsread
Plugins1
apache_airflow.plugins.listGet Pluginsread
Providers1
apache_airflow.providers.listGet Providersread
Task1
apache_airflow.task.getGet Taskread
Task instance dependencies1
apache_airflow.task_instance_dependencies.listGet Task Instance Dependenciesread
Task instance external log URL1
apache_airflow.task_instance_external_log_url.getGet External Log Urlread
Task instance extra links1
apache_airflow.task_instance_extra_links.listGet Extra Linksread
Task instance log1
apache_airflow.task_instance_log.getGet Logread
Task instance tries1
apache_airflow.task_instance_tries.listGet Task Instance Triesread
Task instance try1
apache_airflow.task_instance_try.getGet Task Instance Try Detailsread
Tasks1
apache_airflow.tasks.listGet Tasksread
Version1
apache_airflow.version.getGet Versionread
Xcom entries1
apache_airflow.xcom_entries.listGet Xcom Entriesread
Questions
The Apache Airflow MCP server, answered.
Is there an Apache Airflow MCP server for Claude?
Yes. actionlake is an MCP gateway: a remote MCP server that gives Claude, or another MCP client, 125 Apache Airflow actions across backfill, connection, dag and dag run, through the same eight tools it uses for every other connector. actionlake is independent; it is not made or endorsed by Apache Airflow.
What do I need to connect Apache Airflow?
Username of the Airflow account actionlake signs in as and Password of that Airflow account. actionlake exchanges it at /auth/token for a JWT. Plus your host. The free call account { op: "connections.requirements", service: "apache_airflow" } lists exactly what to paste.
Can an AI agent delete or change things in Apache Airflow?
Only if its key allows it, and only with an explicit confirmation. 19 Apache Airflow actions are destructive, including delete Dag, delete Connection and bulk Pools. 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 Apache Airflow API limit?
actionlake paces every call to the limits in the Apache Airflow service file: 10 calls a second per connection. Repeated reads are answered from cache (58 Apache Airflow actions can be), which never reaches Apache Airflow.
What does an Apache Airflow action cost?
One upstream call, the only unit actionlake bills. A list that pages bills one call per page (125 Apache Airflow actions page). Cache hits, searches, dry runs and calls Apache Airflow 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 Apache Airflow API does it cover?
The Airflow 3.3.2 stable REST API v2 (OpenAPI 3.1) at your Apache Airflow instance's API base URL, which you enter when connecting. It is built from Apache Airflow's documentation as read on 23 October 2026. 125 actions in 70 areas, each one in the signed service file (version 1.0.0).
Does actionlake keep my Apache Airflow 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 Apache Airflow 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 Apache Airflow today.
Start free - 100,000 calls/month, no card. No signup. Hard caps on every key. Cached answers are free.