# Hookdeck MCP for AI Agents AI Agent Connect > Hookdeck MCP lets you manage and monitor webhooks directly through your AI agent. You can list connections, create sources, and control event routing without leaving your chat interface. It brings your webhook reliability layer into your daily workflow, making it easy to pause traffic, debug failed events, and check metrics in real-time. ## Overview - **Category:** developer-tools - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_3hoCgh0EJctquMbKp8KpWnYvsZAo4mKmhiaSCaIm/ai-agent-connect - **Tags:** webhooks, event-driven, api-integration, event-routing, observability ## Description Hookdeck lets you connect your account to your AI agent to manage your webhook infrastructure through natural conversation. Hookdeck acts as the reliability layer for your event-driven architecture, and this Connector puts that control directly into your chat interface. You can manage connections, toggle traffic, and monitor your sources without leaving your workspace. Instead of hunting through a dashboard to find a specific route during an outage, you can just tell your agent to pause it. You can also audit your setup, check queue depths, and debug failed events by fetching specific attempts or requests. If you need to test a new data mapping, you can run a transformation against a sample payload to see the result before you go live. This Connector makes it easy to keep your data moving while giving you the tools to stop it instantly when things go wrong. You can find this in the Vinkius catalog to get started. ## Tools ### delete_destination Delete a destination from your routing setup. This removes the endpoint from your delivery list. ### disable_source Deactivate a source to prevent it from sending data. This is a high-level way to stop incoming traffic. ### count_connections Get the total number of connections in your project. This helps you keep track of your routing scale. ### create_connection Set up a new connection to route webhooks to a destination. Use this to build out new data paths. ### create_destination Add a new destination where your webhooks will be sent. This defines the final stop for your data. ### create_issue_trigger Create a new trigger to alert you when issues occur. This helps you stay on top of webhook failures. ### create_source Add a new webhook source to your Hookdeck project. Use this to bring in new data providers. ### create_transformation Define a new transformation to modify incoming webhook data. This lets you shape your data on the fly. ### delete_connection Permanently remove a connection from your project. Use this to decommission old data routes. ### delete_issue_trigger Remove an issue trigger from your project. Use this to clean up old alerts. ### delete_source Delete a source from your project configuration. This stops all data flow from a specific provider. ### disable_connection Turn off a connection to stop event routing immediately. Use this as a circuit breaker during incidents. ### get_attempt Retrieve the details of a specific delivery attempt. This shows you exactly what happened during a failure. ### get_event Pull the data and status for a specific webhook event. This is perfect for inspecting individual payloads. ### get_issue_trigger Retrieve the configuration for a specific issue trigger. Use this to check your active alerts. ### get_metrics_transformations See how your data transformations are performing. Use this to check for latency in your data processing. ### get_request Retrieve a specific request from your webhook history. Use this to find the exact payload that caused an error. ### get_transformation View the code and logic of a specific transformation. Use this to audit your data mapping rules. ### list_attempts See a list of recent delivery attempts for your events. This is your first stop for troubleshooting failures. ### list_connections See every connection currently configured in your project. This gives you a full view of your routing. ### list_sources See a list of all webhook sources in your project. This helps you manage your various data integrations. ### list_transformations View all the transformations you've defined. Use this to see how you are shaping your incoming data. ### pause_connection Stop event routing for a specific connection immediately. This is your primary tool for handling outages. ### retry_request Try sending a request that was previously rejected. Use this to re-process failed requests manually. ### test_transformation Run your transformation code against a sample payload to see the output. This prevents errors in production. ### trigger_bookmark Replay a request that you saved in a bookmark. This is perfect for simulating specific edge cases. ### unpause_connection Resume event routing for a connection that was previously paused. Use this to restore traffic flow. ### update_destination Modify the URL or rate limits of a destination. Use this to point your data to new endpoints. ### update_issue_trigger Edit the configuration of an existing issue trigger. Use this to fine-tune your alerts. ### count_sources See how many webhook sources you have active. It is useful for auditing your total integrations. ### create_bookmark Save a specific request to a bookmark for later replay. This is great for testing specific payloads. ### delete_bookmark Remove a saved request from your bookmarks. Use this to clean up your saved test cases. ### enable_connection Turn back on a connection to resume event routing. Use this once your backend issues are resolved. ### enable_source Reactivate a source to begin receiving webhook data. Use this to restore data flow from a provider. ### get_connection Fetch the configuration and status of a specific connection. Use this to check your current routing rules. ### get_destination Get the details of a specific destination endpoint. This shows you where your data is being sent. ### get_metrics_attempts View delivery attempt metrics for your webhooks. This helps you monitor success rates over time. ### get_metrics_events See statistics on how your events are being processed. Use this to monitor your overall traffic volume. ### get_metrics_queue_depth Check how many events are waiting in a destination's queue. Use this to spot bottlenecks early. ### get_metrics_requests Get the volume of incoming requests for your project. This helps you understand your peak traffic times. ### get_source Fetch the details of a specific webhook source. This shows you the configuration of your data providers. ### list_bookmarks View all the requests you've saved as bookmarks. Use this to manage your saved test scenarios. ### list_destinations List all the destinations where your data is sent. Use this to see every active endpoint in your project. ### list_events View a list of recent webhook events. This is useful for quickly scanning recent activity. ### list_issue_triggers See all the issue triggers you've created. Use this to audit your current alerting setup. ### list_requests Pull a list of recent incoming webhook requests. Use this to see the latest data hitting your project. ### retry_event Manually trigger a retry for an event that failed to deliver. Use this to recover data without manual intervention. ### update_connection Change the name or routing rules of an existing connection. Use this to adjust your data paths on the fly. ### update_source Modify the settings of a webhook source. Use this to update credentials or URLs for your providers. ### update_transformation Edit the code for a specific data transformation. Use this to update your data mapping logic. ### upsert_connection Create a new connection or update one by its name. Use this to manage your connections efficiently. ### cancel_event Stop any scheduled retries for an event. Use this when you want to prevent an event from trying to deliver again. ## Prompt Examples **Prompt:** ``` List all my current Hookdeck connections. ``` **Response:** ``` I've retrieved your connections. You have 3 active routes: | Route Name | ID | Status | | :--- | :--- | :--- | | **Shopify-to-Lambda** | `conn_123` | Active | | **Stripe-to-Analytics** | `conn_456` | Paused | | **GitHub-Webhooks** | `conn_789` | Active | ``` **Prompt:** ``` Pause the connection with ID conn_456 immediately. ``` **Response:** ``` The connection **Stripe-to-Analytics** (`conn_456`) has been successfully paused. * **Action:** Event routing suspended. * **Reason:** Manual request. ``` **Prompt:** ``` How many webhook sources do I have? ``` **Response:** ``` You currently have a total of **5 sources** configured in your Hookdeck project. Would you like me to list them for you? ``` ## Capabilities ### Toggle webhook traffic Use pause_connection and unpause_connection to stop or resume event routing instantly. ### Audit connection counts Get a quick overview of your infrastructure using count_connections and count_sources. ### Create new routing rules Set up new paths for your data using create_connection and create_destination. ### Debug failed deliveries See exactly why a webhook failed with get_attempt and retry_event. ### Test data transformations Verify your logic with test_transformation to see the output before shipping code. ### Manage source lifecycle Enable or disable sources using enable_source and disable_source to control data flow. ## Use Cases ### Incident Response During Outages An engineer notices a spike in errors and uses pause_connection to stop the flow immediately while they fix the backend. ### Debugging Failed Payloads A developer uses get_request to see the exact payload that caused a transformation to fail and then uses retry_event to try again. ### Infrastructure Auditing A specialist uses list_connections to audit all active routes before moving to a new destination or decommissioning old ones. ### Pre-deployment Testing A QA engineer uses test_transformation to verify that a new order_paid webhook correctly formats the price before going live. ## Benefits - Stop manual dashboard clicking by using pause_connection to toggle traffic in seconds during an incident. - Debug failed deliveries faster with get_attempt and retry_event to see exactly why a webhook died and how to fix it. - Test your logic before shipping with test_transformation to ensure your data mapping is correct before it hits production. - Audit your entire infrastructure at a glance using count_connections and list_sources to keep your project organized. - Replay specific scenarios using trigger_bookmark to simulate edge cases without having to re-trigger the original source. - Monitor health in real-time with get_metrics_queue_depth to spot bottlenecks before they crash your downstream systems. ## How It Works The bottom line is you get a hands-on way to control your webhook traffic and debug events without leaving your chat. 1. Subscribe to the Hookdeck MCP on Vinkius. 2. Enter your Hookdeck API Key into your AI client configuration. 3. Ask your agent to list connections or pause specific webhook routes. ## Frequently Asked Questions **How does the Hookdeck MCP help with webhook reliability?** It gives you a way to manage your webhook infrastructure through your AI agent. You can pause traffic to prevent system overload, manage retries, and monitor your routing rules in real-time. **Can I use Hookdeck MCP to debug failed webhooks?** Yes. You can ask your agent to fetch specific delivery attempts or inspect the exact payloads of failed requests to see exactly what went wrong. **How do I pause a webhook with the Hookdeck MCP?** Just tell your agent to pause a specific connection by name or ID. It will immediately stop the event routing for that route, acting as a circuit breaker. **Can the Hookdeck MCP manage multiple webhook sources?** Yes, it can manage all of your sources. You can list them, count them, and enable or disable them to control where your data comes from. **How does the Hookdeck MCP handle data transformations?** You can use it to view, update, and test your data transformations. This ensures your data is shaped correctly before it reaches your destination. **Can I use the Hookdeck MCP to replay old requests?** Yes, by using the bookmarking feature. You can save specific requests and then ask your agent to replay them whenever you need to test a scenario. **Does Hookdeck MCP work with Claude or Cursor?** Yes, it works with any MCP-compatible client, including Claude, Cursor, and Windsurf, as long as you have the Hookdeck MCP connected. **How do I see my webhook queue depth with Hookdeck MCP?** You can ask your agent to check the queue depth for any destination. This helps you see if your events are backing up and need more processing power. **How can I temporarily stop webhooks from being delivered to a destination?** You can use the `pause_connection` tool with the specific Connection ID. This will halt event routing until you call `unpause_connection`. **Can I see how many active sources I have configured?** Yes, use the `count_sources` tool to get the total number of sources, or `list_sources` to see the details of each one. **Is it possible to update connection rules through the AI?** Absolutely. Use the `update_connection` tool by providing the Connection ID and a JSON body containing the new configuration rules.