Dune Analytics MCP. Query Web3 SQL Metrics from Your Agent
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
Dune Analytics (Web3 SQL Analytics API) lets your AI agent run SQL queries directly against live blockchain data. You can track executions, get status updates, and pull large result sets across Ethereum, Solana, and Polygon.
Stop switching between terminals and dashboards; analyze Web3 metrics from your chat interface.
What your AI agents can do
Cancel execution
Stops a query that is currently running or stuck.
Execute query
Starts a specific Dune query ID and returns an ID you use to track its status.
Get execution results
Retrieves the actual data rows from a query execution, but only if the job is finished.
Execute a specific Dune query ID using your agent, returning an ID you must track.
Check the current state of a running query execution using its ID.
Pull the final data rows from a query only after the status confirms completion.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
Dune Analytics (Web3 SQL Analytics API): 4 Tools for Web3 Data
Manage the entire lifecycle of blockchain data queries. Use these tools to execute, track, cancel, and retrieve results from Dune Analytics.
019e5d14cancel execution
Stops a query that is currently running or stuck.
019e5d14execute query
Starts a specific Dune query ID and returns an ID you use to track its status.
019e5d14get execution results
Retrieves the actual data rows from a query execution, but only if the job is finished.
019e5d14get execution status
Checks the current state of a query execution (e.g., pending, completed, failed).
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 Dune Analytics (Web3 SQL Analytics API), 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
Dune Analytics lets your AI agent run SQL queries right against live blockchain data. You're dealing with Ethereum, Solana, and Polygon data, so you can stop jumping between dashboards and terminals. You just talk to your agent, and it handles the Web3 metrics.
To get started, your agent uses the execute_query tool. You feed it a specific Dune query ID, and it kicks off the query, giving you an execution ID that you've gotta track. You'll use this ID to check the status with get_execution_status. That tells you if the query is pending, completed, or if it failed.
Once the status says 'completed,' you can pull the actual data rows using get_execution_results. If a query runs too long or you change your mind, you can kill it with cancel_execution. This keeps your API usage clean.
When you're done, you've got all the final data right there in your chat window. You can analyze it, visualize it, or build a report immediately. You're running complex SQL against massive, live datasets, and you're doing it without leaving your chat interface. This is how you analyze Web3 metrics.
How Dune Analytics MCP Works
- 1 First, your agent calls
execute_querywith the Dune query ID and any required parameters. This starts the job and returns anexecution_id. - 2 Next, your agent uses
get_execution_statusrepeatedly, passing theexecution_id, until the status returnsQUERY_STATE_COMPLETED. - 3 Finally, the agent calls
get_execution_resultswith theexecution_idto pull the actual data rows for analysis.
The bottom line is, you use the tools together to manage the state of a long-running job, ensuring you only pull data when the job is actually done.
Who Is Dune Analytics MCP For?
This is for data analysts, web3 developers, and crypto researchers who need raw, verifiable on-chain data. If your job involves checking token movements, calculating DeFi metrics, or tracking protocol health, this server is necessary. It lets you move data analysis from dedicated dashboards into your everyday workspace.
Pulls the latest DeFi, NFT, or DAO metrics by running specific queries without leaving their analysis environment.
Verifies on-chain states, checks protocol health, or validates contract interactions directly from their code editor or agent.
Automates the gathering of ecosystem growth data and market trends needed for reports, managing complex, multi-step data gathering.
What Changes When You Connect
- See real-time DeFi, NFT, or DAO metrics by calling
execute_querywith a specific Dune query ID. This gives you immediate access to on-chain data without leaving your chat or IDE. - Keep track of long-running jobs by using
get_execution_status. You get the current state (pending, running, failed) and the execution ID, letting you know exactly what's happening. - Fetch large result sets directly into your agent using
get_execution_results. You get the raw data rows ready for visualization or report writing. - Control your API usage by calling
cancel_execution. You get the ability to stop expensive or unnecessary queries before they run too long. - Analyze multiple chains (Ethereum, Solana, Polygon) from one place. You get a unified interface for Web3 data that used to require multiple specialized dashboards.
Real-World Use Cases
Checking Protocol Health
A developer needs to confirm the total TVL of a new protocol. They ask their agent to run a specific Dune query ID. The agent uses execute_query to start the job, monitors it with get_execution_status, and finally calls get_execution_results to get the precise, up-to-date number for the report.
Tracking Market Trends
A researcher needs to gather market activity across several DAOs. They ask their agent to run three different queries, managing each one's state sequentially. If one query stalls, they can call cancel_execution to save credits and then retry the job.
Validating Transaction Data
An analyst wants to verify the top 5 token holders. They use their agent to run a complex query ID. The agent uses get_execution_status to wait for completion and then pulls the results with get_execution_results to summarize the holdings.
Stopping a Failed Query
A complex query starts running but is found to be based on outdated parameters. Instead of waiting, the analyst tells their agent to call cancel_execution. This immediately halts the process, saving time and API credits.
The Tradeoffs
Treating it like a single API call
The developer assumes calling execute_query immediately provides the final data, or that the result will stream back instantly, causing the agent to block or time out.
→
The job is asynchronous. First, use execute_query to get the ID. Then, poll the status using get_execution_status until the status is QUERY_STATE_COMPLETED. Only then, call get_execution_results.
Ignoring the state lifecycle
The user runs a query and immediately tries to fetch the data using get_execution_results before the status is complete, leading to an error or empty result set.
→
Always check the state first. Use get_execution_status to confirm the status is COMPLETED. If it's PENDING, wait. If it's FAILED, review the error message.
Not cancelling abandoned jobs
A long-running query is started but is no longer relevant. The user leaves the tab open, wasting API credits and slowing down the account.
→
Use cancel_execution immediately if the query is unnecessary or taking too long. This sends a direct command to halt the job.
When It Fits, When It Doesn't
Use this server if your workflow requires managing long-running, asynchronous data tasks against external, structured data sources like blockchain records. You need to reliably check the status of a job before you can retrieve the data. If you only need to read static, in-memory data, or if your data source is simple and synchronous, don't use this. You'll waste time calling get_execution_status just to find out the job failed anyway. This tool is for stateful data pipelines.
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Dune Analytics. 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 4 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Pulling Web3 data used to mean switching between 3-4 dashboards.
Today, getting a clear picture of a DeFi metric requires hopping between the Dune dashboard, a terminal for SQL, and a spreadsheet to summarize the results. You're copy-pasting IDs, refreshing tabs, and manually correlating data points just to answer a simple question about token distribution.
With the Dune Analytics MCP Server, you talk to your agent. You give it a query ID, and it handles the entire process—starting the job, waiting for the result, and pulling the raw data into your chat. It's all one conversation.
Dune Analytics MCP Server: Get the data, not just the query.
The old way meant running a query and then having to manually figure out if the job succeeded, what the error was, and then going back to another interface to download the results. It was a multi-step, manual process with no single source of truth.
Now, the agent manages the entire state machine. It calls `get_execution_status` until it's done, and only then calls `get_execution_results`. You get the final, structured data right where you are working.
Common Questions About Dune Analytics MCP
How do I start a query with the Dune Analytics (Web3 SQL Analytics API) MCP Server? +
You start by calling execute_query with the specific Dune query ID and any parameters. This tells the system to start the job and returns an execution ID.
What if my query fails when using the Dune Analytics (Web3 SQL Analytics API) MCP Server? +
You check the status using get_execution_status. If the status is QUERY_STATE_FAILED, the system provides details on why the query failed. You then adjust your parameters and try again.
Can I cancel a query running on the Dune Analytics (Web3 SQL Analytics API) MCP Server? +
Yes, you call cancel_execution with the execution ID. This immediately stops the query, saving your API credits and letting you troubleshoot the job.
Do I need to run `get_execution_results` after checking status? +
Yes. get_execution_results pulls the actual data. You must confirm the status is COMPLETED first using get_execution_status; otherwise, the result will be empty.
How do I manage the parameters when I use the `execute_query` tool in the Dune Analytics (Web3 SQL Analytics API) MCP Server? +
You pass the parameters directly when calling execute_query. These parameters allow you to customize the query, like setting specific date ranges or contract addresses, ensuring you get the precise data you need.
What happens if I try to fetch results using `get_execution_results` before the query is finished? +
The API will return an error or an incomplete status. Always use get_execution_status first. This confirms the query has reached QUERY_STATE_COMPLETED before you attempt data retrieval.
Does the Dune Analytics (Web3 SQL Analytics API) MCP Server handle large data sets effectively? +
Yes. The get_execution_results tool is designed to retrieve large result sets. Your agent handles the data stream, allowing for immediate analysis or chunking for visualization.
Is there a specific setup required for the Dune Analytics (Web3 SQL Analytics API) MCP Server? +
You must subscribe to the server and provide your personal Dune API Key. Once authenticated, your AI agent can access the tools and begin querying blockchain data.
How do I retrieve the actual data rows from a query? +
First, use execute_query to start the process and get an execution_id. Once get_execution_status shows the state is completed, use get_execution_results with that ID to fetch the data.
Can I stop a query that is taking too long to run? +
Yes. You can use the cancel_execution tool by providing the active execution_id. This will immediately interrupt the ongoing query on Dune's servers.
How do I pass parameters like dates or addresses to my Dune query? +
When calling execute_query, use the query_parameters field. It accepts a JSON string (e.g., '{"address": "0x...", "limit": 10}'). The agent will parse this and apply it to your SQL execution.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Nasdaq Data Link (Quandl)
Access financial, economic, and alternative data from Nasdaq Data Link. Query datatables, fetch metadata, and manage bulk downloads directly from your AI agent.
Fairing
Analyze customer insights via Fairing — manage post-purchase surveys, track responses, and query zero-party data through your AI agent.
IBGE Pesquisas — Brasil Cidades
Query the engine behind Brasil Cidades: health, education, economy, and quality-of-life indicators for every Brazilian municipality — rank cities, compare regions, and build data-driven policy analysis.
You might also like
Fastn
Automate and execute low-code workflows via Fastn — manage flow definitions, monitor executions, and handle connector credentials directly.
MintMCP
Enterprise MCP Gateway: manage authentication, monitoring, and security guardrails via centralized virtual servers.
Markdown Task Extractor
Instantly scan any local folder of Markdown notes (Obsidian, Notion, Logseq) and aggregate every scattered '- [ ]' to-do item into a single, structured list for your AI.