Goldsky MCP. Manage Web3 indexing pipelines from your 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.
Goldsky (Web3 Data Indexing & Subgraphs) MCP Server manages high-performance Web3 data pipelines. Use this to list, create, pause, restart, and monitor blockchain data indexes directly from your AI agent.
Check status, retrieve execution logs, and validate definitions without leaving your editor.
What your AI agents can do
Create pipeline
Deploys a new Web3 data pipeline based on a defined JSON structure.
Delete pipeline
Removes a specified pipeline definition from the system (this action is irreversible).
Get pipeline
Retrieves the detailed configuration and metadata for a single, specific pipeline.
See all existing pipelines and deploy brand new ones using a structured definition.
Pause, resume, or restart any pipeline to control resource consumption or force an update.
Check a pipeline's current operational status, including whether it's running, paused, or failed.
Retrieve detailed execution logs and error counts for specific time windows to troubleshoot data flow issues.
Test a pipeline's source, transform, and sink definitions before deployment to catch configuration errors.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Goldsky (Web3 Data Indexing & Subgraphs) MCP Server: 12 Tools
Use these 12 tools to control the entire lifecycle of your blockchain data indexing pipelines, from initial validation to full monitoring.
019e5d20create pipeline
Deploys a new Web3 data pipeline based on a defined JSON structure.
019e5d20delete pipeline
Removes a specified pipeline definition from the system (this action is irreversible).
019e5d20get pipeline
Retrieves the detailed configuration and metadata for a single, specific pipeline.
019e5d20get pipeline error count
Checks the number of errors that occurred in a pipeline within a specified time window.
019e5d20get pipeline logs
Fetches the sequence of execution logs for a chosen pipeline.
019e5d20get pipeline state
Gets the internal, detailed state of a pipeline's current operation.
019e5d20get pipeline status
Reports the current, high-level operational status of a pipeline (e.g., running, paused, failed).
019e5d20list pipelines
Provides a complete list of all data pipelines configured in the project.
019e5d20pause pipeline
Temporarily halts a running pipeline to manage resources or apply updates.
019e5d20restart pipeline
Forces a pipeline to shut down and then immediately start running again.
019e5d20resume pipeline
Brings a pipeline back to an active state after it was previously paused.
019e5d20validate pipeline
Checks a pipeline definition against schema rules before attempting deployment.
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 Goldsky (Web3 Data Indexing & Subgraphs), 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
Goldsky (Web3 Data Indexing & Subgraphs) manages your high-performance Web3 data pipelines. You can use this server to list, create, pause, restart, and monitor blockchain data indexes directly from your AI agent. You can check the status, pull execution logs, and validate definitions without ever leaving your editor.
List and Create Pipelines
You can see every pipeline you've got running with list_pipelines and deploy a brand new one using create_pipeline with a structured JSON definition. You can also remove a pipeline definition completely with delete_pipeline, and remember, that action's irreversible.
Control Pipeline State
If you need to manage resources or push an update, you can temporarily stop a running pipeline using pause_pipeline. You can bring it back to life with resume_pipeline, or you can force it to shut down and start up again immediately by calling restart_pipeline.
Monitor Execution Status
To see if a pipeline is running, paused, or failed, just check its high-level operational status with get_pipeline_status. For a deeper dive, you can pull the detailed internal state of the pipeline's current operation using get_pipeline_state.
Debug and Audit Logs
Troubleshooting data flow issues is easy. You can pull the full sequence of execution logs for a chosen pipeline using get_pipeline_logs, and you can check how many errors popped up in a specific time window with get_pipeline_error_count.
Validate Definitions
Before you even try deploying something, you can check a pipeline definition's schema rules using validate_pipeline. This catches configuration errors for your sources, transforms, and sinks before they break production.
How Goldsky MCP Works
- 1 First, connect the Goldsky MCP Server to your AI client and provide your API key.
- 2 Then, issue a natural language command, like 'List all active pipelines' or 'Pause the Polygon NFT events pipeline'.
- 3 The server executes the corresponding tool call, and your AI client returns the real-time status or data to you.
The bottom line is: you manage your complex, mission-critical Web3 data infrastructure entirely through chat commands.
Who Is Goldsky MCP For?
Web3 Developers and Data Engineers need this. If you spend time debugging why a blockchain index stopped flowing or manually checking deployment readiness across multiple dashboards, this server saves you time. It lets you run full pipeline lifecycle management—from listing to logging—without leaving your code editor.
Deploys, debugs, and debugs subgraphs and data pipelines directly from their IDE using conversational commands.
Monitors indexing health, retrieves execution logs, and checks error counts to troubleshoot data flow issues in real-time.
Manages cross-chain data indexing infrastructure and coordinates pipeline status checks across multiple services.
What Changes When You Connect
- Full Lifecycle Control: You can pause, resume, or restart pipelines using
pause_pipeline,resume_pipeline, andrestart_pipeline. This gives you granular control over resource usage without logging into a separate dashboard. - Instant Debugging: Instead of digging through multiple dashboards, use
get_pipeline_logsto pull recent execution logs right into your chat. You see exactly what happened when the data flow broke. - Pre-Deployment Safety: Before deploying any new index, run
validate_pipeline. This tool checks the definitions (sources, transforms, sinks) and stops you from deploying bad configuration. - Auditability: Use
get_pipeline_error_countto query error metrics across specific time windows. This is better than just looking at a 'last error' counter—it gives you a true health picture. - Visibility: The
list_pipelinestool gives you an immediate overview of every pipeline in the project. You know exactly what's running and what's dormant. - Deep State Inspection: Need to know why a pipeline failed?
get_pipeline_stategives you the internal machinery status, going deeper than a simple 'failed' status.
Real-World Use Cases
Investigating a sudden data drop
A data engineer notices the Polygon NFT events pipeline stopped indexing. They ask their agent to run get_pipeline_status (it's 'failed'), then use get_pipeline_logs to pull the last 50 lines. They see a connection timeout, fix the source definition, and use create_pipeline to redeploy it. Problem solved.
Preparing for a cross-chain launch
A protocol team needs to add a new Base bridge monitor. First, they use list_pipelines to see current indexes. Then, they use validate_pipeline to ensure their new definition is sound. Finally, they use create_pipeline to launch the index, all without leaving their command line.
Scaling up a critical index
The swaps pipeline is hitting rate limits. The developer pauses it with pause_pipeline. They then update the underlying source connection, and finally resume it with resume_pipeline. This controlled process prevents resource exhaustion.
Debugging a complex transformation error
The data flow is producing bad data. The engineer uses get_pipeline_error_count to confirm the error spike started yesterday. They then use get_pipeline_state to check the internal data sink connection and find the source of the issue.
The Tradeoffs
Checking status in a loop
Manually running get_pipeline_status every 10 seconds in a script until it says 'running'. This is slow, wasteful, and hits rate limits.
→
Instead, use the AI agent to query the status once and then rely on the agent's built-in notification features, or use get_pipeline_state for a single, deep check.
Forgetting to validate new definitions
A developer skips running a pre-deployment check and attempts to run create_pipeline with a new, malformed JSON definition. The deployment fails hours later, requiring manual rollback.
→
Always run validate_pipeline first. It confirms the definitions are sound before you spend time creating the pipeline.
Ignoring pipeline dependencies
A developer tries to resume a pipeline that depends on a source connection that was deleted. The operation fails, and they spend time figuring out why the failure happened.
→
Always use get_pipeline first to confirm the configuration details and ensure all required components are present.
When It Fits, When It Doesn't
Use this MCP Server if your primary job is managing the full lifecycle of Web3 data indexes. This is for Web3 Developers and Data Engineers who need to go beyond simple reads and control the infrastructure itself. You need tools like create_pipeline, pause_pipeline, and get_pipeline_logs.
Don't use this if you just need to read a single piece of data. For simple data queries, use a standard graph query tool. If you only need to monitor health, get_pipeline_status might be enough, but if you need to fix the problem, you need the full suite of tools here.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Goldsky. 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 12 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Debugging data pipelines used to mean bouncing between dashboards.
Today, if your blockchain index breaks, you open the Goldsky dashboard. You check the status panel. If it failed, you copy the error ID. You then switch to the logs tab, paste the ID, and filter by time. If you want to fix it, you find the source definition, adjust the JSON, and then try to click 'run' again. It’s a multi-tab, copy-paste, context-switching nightmare.
With the Goldsky MCP Server, you just tell your agent, 'Check the swaps pipeline and show me the logs.' The agent runs `get_pipeline_status` and `get_pipeline_logs` in sequence, pulls the data, and hands it back to you in the chat. The whole debugging flow stays in one place.
The Goldsky MCP Server: Full Pipeline Management
Before this server, managing a pipeline update meant: 1) Manually pausing the service, 2) Copying the old configuration, 3) Editing the JSON locally, and 4) Restarting the service, hoping the change stuck.
Now, you tell your agent to `pause_pipeline`, give it the updated definition, and ask it to `create_pipeline`. The server handles the whole sequence, and you get the confirmation without touching a dashboard.
Common Questions About Goldsky MCP
How do I check the health of my pipelines using get_pipeline_status? +
It reports the current operational state (running, paused, failed). For a quick check, ask your agent to run get_pipeline_status on the pipeline name. If it says 'failed', you'll need to run get_pipeline_logs next.
Can I deploy a new pipeline using create_pipeline? +
Yes. You provide the full JSON definition, and the server handles the deployment. Remember that validate_pipeline should always run first to make sure your definition is correct.
What is the difference between get_pipeline_state and get_pipeline_status? +
get_pipeline_status gives you the high-level status (running, paused). get_pipeline_state gives you the deeper, internal machine state—it's more detailed and useful for advanced debugging.
How do I delete a pipeline using delete_pipeline? +
Just ask your agent to run delete_pipeline and confirm the pipeline name. Be careful, this action is irreversible.
How do I check for data flow errors using get_pipeline_error_count? +
It calculates the number of errors within a specified time window. You pass the pipeline name and the time range to get a precise count, helping you track data indexing health over time.
What tool do I use to validate a new pipeline definition before deployment? (validate_pipeline) +
Use the validate_pipeline tool. It checks the definition—sources, transforms, and sinks—for configuration errors, preventing you from deploying bad code.
How can I manage a paused pipeline using resume_pipeline? +
resume_pipeline reactivates a pipeline that was previously paused. You just provide the pipeline name, and it starts processing data again.
What is the difference between list_pipelines and get_pipeline? +
list_pipelines shows all active pipelines in the project. get_pipeline retrieves specific, detailed information about one pipeline by name.
Can I check if my pipeline configuration is valid before deploying it? +
Yes! Use the validate_pipeline tool with your definition object. It will check your sources, transforms, and sinks for errors without actually starting a new pipeline.
How do I monitor errors in a running pipeline? +
You can use get_pipeline_error_count to see the number of issues in a time window, or get_pipeline_logs to fetch the actual execution logs for detailed debugging.
Is it possible to temporarily stop a pipeline without deleting it? +
Absolutely. Use the pause_pipeline tool to stop execution. When you are ready to start again, simply use the resume_pipeline tool.
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.
Deterministic Codec Engine
Empower your AI to perfectly serialize and deserialize data. Effortlessly switch between URL Encoding, HTML Entities, Unicode Escapes, and DNS Punycode with a native V8 engine.
Modelbit (ML Model Deployments)
Deploy and call machine learning models directly from your AI agent using Modelbit's inference endpoints.
You might also like
Smartsheet
Empower your AI to read Smartsheet rows, list workspaces, and manage your spreadsheets effortlessly from your code editor.
SendPulse
Manage email marketing and automation via SendPulse — handle mailing lists, manage contacts, and track campaigns directly from your AI agent.
Ghostfolio (Investment Tracker)
Track your wealth and manage investment portfolios via Ghostfolio — monitor holdings, record activities, and analyze performance through AI.