Dagster MCP. Audit your data pipelines and assets from chat.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Dagster connects your data orchestration layer to your AI agent. You use it to manage data pipelines, audit job schedules, and track software-defined assets from natural conversation.
It lets you list jobs, check run histories, and verify data dependencies without leaving your IDE or chat window. It’s full data governance for data engineers.
What your AI agents can do
Get run
Retrieves detailed status and execution logs for a specific job run ID.
List assets
Lists all software-defined assets managed by the Dagster instance.
List jobs
Lists all data jobs defined in the Dagster environment.
Use list_jobs to retrieve a list of all data jobs defined in your Dagster deployment.
Use list_runs to fetch a list of recent job execution runs and check their status.
Use get_run to retrieve the specific status and logs for a single, identified job run.
Use list_assets to enumerate all software-defined assets and map their dependencies.
Use list_schedules to list all configured job schedules for auditing purposes.
Use list_sensors to list all active sensors that listen for external events.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Dagster MCP Server: 6 Tools for Data Pipelines
Manage data pipelines, track job runs, and audit asset states by calling these 6 specialized tools from your AI agent.
019d7581get run
Retrieves detailed status and execution logs for a specific job run ID.
019d7581list assets
Lists all software-defined assets managed by the Dagster instance.
019d7581list jobs
Lists all data jobs defined in the Dagster environment.
019d7581list runs
Retrieves a list of recent job runs and their high-level statuses.
019d7581list schedules
Lists all configured job schedules that run automatically.
019d7581list sensors
Lists all active sensors that listen for external events to trigger data actions.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Dagster, then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
Dagster MCP Server - Manage Data Pipelines
Your AI client talks directly to your Dagster data orchestration layer. You use this server to audit data pipelines, track assets, and manage job schedules using natural conversation. It's full data governance for data engineers, letting your agent list jobs, check run histories, and verify data dependencies without you leaving your chat window or IDE.
Audit and list active jobs
Use list_jobs to grab a list of every data job defined in your Dagster deployment.
Review job run history
Run list_runs to fetch a list of recent job executions and see their high-level status. You can then use get_run to get the specific status and full logs for any single, identified job run.
Check data asset dependencieslist_assets enumerates every software-defined asset and maps out all their dependencies.
View automation triggerslist_schedules lists all configured job schedules that run automatically. You can also check for external event listeners using list_sensors, which lists every active sensor listening for external events.
Monitor the systemlist_jobs lets you see what data jobs are active. list_runs gives you a quick look at recent job runs and their statuses. get_run pulls the detailed status and execution logs for a specific job run ID. list_assets lets you count all the software-defined assets and map their dependencies. list_schedules shows you all the scheduled jobs. list_sensors lists all the sensors that listen for outside events.
How Dagster MCP Works
- 1 Subscribe to the server and provide your Dagster URL and API token.
- 2 Your AI client sends a natural language request (e.g., 'Show me failed runs from yesterday').
- 3 The agent calls the appropriate tool (e.g.,
list_runs) and returns the structured data to your client.
The bottom line is, you manage your entire data stack—from raw assets to scheduled jobs—using natural conversation.
Who Is Dagster MCP For?
This is for the data platform team, the data engineer, and the analytics engineer. You're the person who wakes up at 2 a.m. because the daily ETL job failed, and you're tired of clicking through three different dashboards just to find the error log. You need to audit pipeline health instantly, right where you're working.
Uses the agent to monitor pipeline health, identify failed runs using list_runs, and pinpoint the specific error log via get_run without leaving their IDE.
Runs list_assets to track data dependencies and verify data freshness in real-time when building a new reporting view.
Audits job schedules and sensor configurations using list_schedules and list_sensors across multiple organizational deployments.
What Changes When You Connect
- Check the full status of data assets using
list_assets. You instantly see the lineage and dependencies between raw data and final reports. - Pinpoint failure sources quickly. Instead of searching dashboards, use
list_runsto get recent job history, and thenget_runto pull the exact error logs. - Audit automated triggers. Use
list_schedulesandlist_sensorsto verify every job that runs automatically, ensuring your data flow is predictable. - Understand your entire codebase. The
list_jobstool shows every defined pipeline boundary, giving you a full map of what runs on your server. - Keep the process in context. You manage job runs, assets, and schedules all through your chat interface, eliminating context switching between tools.
- Verify deployment state. Use the server to identify deployment boundaries and confirm connectivity across your Dagster cluster.
Real-World Use Cases
Debugging a broken pipeline run
A data engineer notices a job failed. They ask their agent to use list_runs to find the run ID, then use get_run to pull the full logs and determine if the failure was due to bad input data or a code bug. Problem solved in chat.
Mapping data dependencies
An analytics engineer needs to know if a new report depends on a specific raw data source. They run list_assets to enumerate all software-defined assets, instantly mapping the dependency chain from source to view.
Auditing scheduled jobs
A data platform manager needs to verify if a critical job runs every hour. They use list_schedules to check the configured frequency and then use list_jobs to confirm the job name, ensuring compliance.
Checking for external triggers
A developer needs to know which data flow responds to a file drop. They use list_sensors to list active sensors and confirm which external events are currently triggering data ingestion.
The Tradeoffs
Searching manually in the UI
You open the Dagster UI, click on 'Runs,' filter by date, then click on 'Assets' to check dependencies, and copy/paste IDs between tabs.
→
Instead, ask your agent to run list_runs for the date range, then reference the asset IDs you see using list_assets within the same chat session. It keeps the context alive.
Assuming all data is visible
You think a job runs because it should run, but you never checked if it's scheduled or if the sensor is active.
→
Always check the triggers. Use list_schedules for time-based runs, and list_sensors for event-based runs. This verifies the actual automation boundaries.
Trying to debug without IDs
You know something is wrong with a job, but you don't have the specific run ID or job ID handy.
→
First, use list_jobs to get the name, then use list_runs to get the recent ID. Finally, use get_run with the ID to get the detailed failure log.
When It Fits, When It Doesn't
Use this if you need to treat your data orchestration layer like a state machine that can be queried via natural language. You need to audit the why and what of your data flows—i.e., 'Why did this run fail?' or 'What assets depend on this raw table?'. You must use it when your workflow requires checking dependencies (list_assets), auditing triggers (list_schedules, list_sensors), or retrieving detailed run history (get_run).
Don't use this if your goal is simply to deploy code or write a new job definition. For that, you still need to use the Dagster UI or code. This server is purely for observation and management; it's a read-only interface for your data platform's state.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Dagster. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 6 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Debugging data pipelines usually means jumping between five different dashboards.
Right now, finding out why a job failed is a multi-tab ordeal. You jump to the Jobs dashboard to see the run status. Then you click on the specific run ID to get logs. If you need to know what assets were impacted, you have to leave the run details page and go to the Assets tab. You copy-paste IDs and switch contexts until you find the root cause.
With the Dagster MCP Server, you just ask your agent: 'What failed yesterday?' It runs the necessary tools (`list_runs`, `get_run`) and gives you the full context, including the associated asset IDs, all in one chat window. It keeps the context exactly where it belongs.
The Dagster MCP Server gives you full control over `list_assets` and run history.
Manual asset tracking involves running ad-hoc queries against the data catalog, cross-referencing schema definitions, and manually mapping dependencies between source tables and materialized views. It's slow, and it’s easy to miss a link.
Now, your agent runs `list_assets` and shows you the entire dependency graph. You can confirm exactly which assets rely on a specific raw table, immediately knowing the blast radius of any change. It turns a guessing game into a verifiable audit.
Common Questions About Dagster MCP
How do I check the status of a specific run using the Dagster MCP Server? +
Use the get_run tool. You just need to provide the unique run ID. The tool returns the full status, execution logs, and detailed metadata for that specific job execution.
What is the difference between `list_jobs` and `list_runs` in Dagster? +
list_jobs gives you the definition of the job (the code and configuration). list_runs gives you the history of when that job actually executed and what the status was.
Can I list all data assets with the Dagster MCP Server? +
Yes, use list_assets. This tool enumerates all software-defined assets, showing you every data entity defined in your pipeline, regardless of its source.
How do I audit scheduled jobs using `list_schedules`? +
Run list_schedules. This tool displays all configured job schedules, letting you verify the frequency and recurrence pattern of your automated data workflows.
What if a job fails? Can I see the error logs using `list_runs`? +
list_runs only gives a high-level status (Success/Failure). You must use get_run with the specific run ID to retrieve the detailed error logs and full execution history.
How do I check which events trigger jobs using `list_sensors`? +
The list_sensors tool lists all active sensors configured in your deployment. Each sensor specifies the external event type it listens for (e.g., S3 bucket change, database update). This helps you verify what triggers your automation.
What is the difference between `list_jobs` and `list_runs`? +
The list_jobs tool lists the definitions of all data jobs available in your system. In contrast, list_runs provides a chronological history of actual job executions, showing status and timeframes.
How do I find out which assets depend on a specific data source using `list_assets`? +
The list_assets tool enumerates every software-defined asset. By checking the asset metadata, you can identify the dependencies and the physical storage mapping for any given data component.
Can my agent list all software-defined assets in Dagster? +
Yes. Use the 'list_assets' tool. Your agent will retrieve all software-defined assets, allowing you to identify data dependencies and verify physical storage mappings within your pipelines.
How do I check the status of a specific job run? +
Provide the 'run_id' to the 'get_run' tool. Your agent will fetch detailed information for that specific execution, including status (Success, Failure, In Progress) and detailed execution logs.
Can I see active sensors and schedules via the agent? +
Absolutely. Use the 'list_schedules' and 'list_sensors' tools. Your agent will pull the active automation triggers, allowing you to audit which jobs are scheduled and which sensors are listening for external events.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Supabase
Connect your AI to Supabase. Execute database queries, manage users, and trigger PostgreSQL functions directly from your terminal.
Traefik Proxy
Monitor and manage your Traefik Proxy infrastructure — inspect routers, services, and middlewares directly from your AI agent.
Checkly
Manage application monitoring and E2E testing via Checkly — track API uptime, trigger checks, and monitor performance directly from any AI agent.
You might also like
ReqBin
Test and verify API requests using ReqBin Echo — send JSON payloads and inspect the echoed response to debug request structures.
Fusebill (Enterprise Subscription Billing API)
Automate enterprise subscription billing via Fusebill — manage customers, track subscriptions, and handle billing lifecycles directly from any AI agent.
Focus Varejo
Emit NFC-e, NFCom, manifest documents and manage retail invoices via Focus NFe API.