# Convoy MCP for AI Agents AI Agent Connect > Convoy lets you manage your entire webhook infrastructure through your AI agent. You can create endpoints, broadcast events, and monitor delivery status without leaving your workspace. It handles the heavy lifting of webhook routing and lifecycle management, giving you a direct line to your event pipeline. ## Overview - **Category:** developer-tools - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_pACxCKSpV9m8EHUh3B3D0acuTTjU5hz9mILWeJhz/ai-agent-connect - **Tags:** webhooks, api-management, event-delivery, infrastructure, webhook-gateway ## Description Managing webhooks usually means jumping between tabs, checking logs, and manually updating URLs. It is a friction-filled process that slows down deployment cycles and creates room for human error. Convoy changes that by bringing your webhook operations directly into your conversation with your AI agent. Instead of hunting for a specific dashboard to toggle a production endpoint or checking why a specific delivery failed, you just ask your agent to do it. You can build out your infrastructure, manage subscriptions, and handle bulk onboarding in seconds. This Connector is one of the many tools available in the Vinkius catalog, making it easy to plug into your existing workflow without extra overhead. It is about moving from manual configuration to natural language commands, letting you focus on building features while the agent handles the plumbing of your event-driven architecture. You can manage the entire lifecycle of your webhooks, from initial creation to final delivery monitoring, through a single interface. This replaces the need for constant context switching between your code and your infrastructure management tools. ## Tools ### bulk_create_filters Create multiple filters for a subscription in one go to save time. This simplifies complex routing setups. ### create_event_type Define a new category of events for your webhook system. This helps organize your event types for better filtering. ### import_event_types Pull in your event definitions directly from an OpenAPI spec. This saves you from manually typing out event schemas. ### activate_endpoint Turn a paused endpoint back on to resume traffic. This is the quickest way to finish your maintenance window. ### batch_retry_event_deliveries Retry multiple failed event deliveries at once to ensure data consistency. This helps clear backlogs without manual intervention. ### broadcast_event Send a single event to every endpoint that matches your criteria. It is the fastest way to reach all matching subscribers. ### bulk_onboard Add multiple endpoints and subscriptions to your project simultaneously. It eliminates the need for repetitive manual entries. ### create_endpoint Set up a new webhook destination with specific project parameters. You can define the URL and metadata in one step. ### create_event Generate a new event to be dispatched across your infrastructure. Use this to trigger specific actions in your pipeline. ### create_filter Set up a specific filter to route events to the right subscribers. This ensures only relevant data reaches your endpoints. ### create_portal_link Generate a link for users to manage their own webhook settings. This empowers your users with self-service options. ### create_source Establish a new source for your webhook events. This defines where your event data originates. ### create_subscription Link a specific endpoint to an event type with custom filters. This is the core of your webhook routing logic. ### delete_endpoint Remove an old or decommissioned webhook destination. Use this to keep your project clean of unused endpoints. ### dynamic_event Create an event that adapts based on the data it carries. This is useful for flexible, data-driven event types. ### fanout_event Send an event to multiple different endpoints at the same time. It is perfect for multi-service notifications. ### get_delivery_attempt Look up the specific details of a single delivery try. Use this to see exactly why a specific packet failed. ### get_endpoint Fetch the current configuration and status of a specific endpoint. This gives you a quick snapshot of your setup. ### list_delivery_attempts See all delivery results for a specific event. Use this to monitor the health of your outgoing notifications. ### list_endpoints View every webhook destination currently active in your project. This gives you a high-level view of your infrastructure. ### list_event_deliveries Check the history of events sent across your system. Use this to audit your event flow over time. ### list_meta_events Review the internal metadata events for your infrastructure. This is helpful for deep debugging of the system. ### list_subscriptions See all active subscriptions and their routing rules. Use this to verify how your data is being distributed. ### pause_endpoint Temporarily stop traffic to an endpoint for maintenance. It prevents your app from receiving data while it is down. ### retry_event_delivery Manually trigger a retry for a specific failed delivery. Use this to fix individual issues without restarting the whole batch. ### update_endpoint Modify the details of an existing webhook destination. Use this to change URLs or metadata on the fly. ## Prompt Examples **Prompt:** ``` Can you show me all the active webhooks for my production project? ``` **Response:** ``` ### Active Webhooks I've retrieved the active endpoints for your project. Here is the current status: | Name | ID | Status | | :--- | :--- | :--- | | **Production Sync** | `end_002` | ✅ Active | | **Analytics Collector** | `end_003` | ✅ Active | | **Staging Webhook** | `end_001` | ⏸️ Paused | You have 3 total endpoints configured. ``` **Prompt:** ``` I need to pause the 'Marketing Sync' webhook for an hour. ``` **Response:** ``` I have successfully paused the **Marketing Sync** webhook (`end_005`). **Status Update:** - **Endpoint:** Marketing Sync - **Action:** Paused - **Reason:** Maintenance Traffic to this destination will be halted until you choose to activate it again. ``` **Prompt:** ``` Send a test 'order_placed' event to all active subscribers. ``` **Response:** ``` ### Broadcast Initiated I am broadcasting the `order_placed` event to all matching subscribers. **Broadcast Summary:** - **Event Type:** `order_placed` - **Payload:** `{"id": 999, "amount": 45.00}` - **Status:** ✅ Dispatched - **Subscribers Reached:** 12 All matching endpoints have received the test payload. ``` ## Capabilities ### Create and manage webhook endpoints Set up new destinations and manage their configurations through natural language. ### Broadcast events to all subscribers Send a single event to every endpoint that matches your criteria instantly. ### Toggle endpoint status for maintenance Pause or activate endpoints to control traffic flow during updates. ### Debug delivery issues with metadata Retrieve detailed info on why a specific delivery failed to reach its target. ### Bulk onboard new subscriptions Add multiple endpoints and subscriptions to your project in one command. ### Retry failed event deliveries Manually or automatically retry failed messages to ensure data consistency. ## Use Cases ### Maintenance Window Toggles An engineer needs to take a service offline for 10 minutes. They ask the agent to `pause_endpoint` for the Analytics webhook to prevent data loss during the restart. ### Rapid Bulk Onboarding A developer just launched a new product and needs to register 50 new webhook URLs. They use `bulk_onboard` to do it in one prompt instead of clicking 50 times. ### Deep Delivery Debugging An SRE notices a spike in errors. They ask the agent to `list_delivery_attempts` for a specific event to see if the issue is a 500 error or a timeout from the provider. ### Automated Event Testing A QA engineer needs to see how the system handles a user_signed_up event. They use `create_event` to fire a test payload and verify the downstream logic. ## Benefits - Stop manual toggling by using `pause_endpoint` and `activate_endpoint` to control traffic during maintenance windows. - Save hours on setup by using `bulk_onboard` to create dozens of subscriptions and endpoints in one command. - Debug faster by using `list_delivery_attempts` to see exactly why a webhook failed to reach its destination. - Simplify event routing by using `import_event_types` to sync your definitions from an OpenAPI spec automatically. - Execute targeted broadcasts using `broadcast_event` to reach every relevant subscriber instantly. - Improve reliability with `batch_retry_event_deliveries` to clear out backlogs of failed messages quickly. ## How It Works The bottom line is you get a hands-free way to manage webhook infrastructure using only natural language. 1. Connect your Convoy API Key and Base URL to the Connector. 2. Describe the endpoint or event you want to create to your agent. 3. Get immediate confirmation and a status summary of the change. ## Frequently Asked Questions **What is Convoy MCP for?** Convoy is an Connector that allows your AI agent to manage your webhook infrastructure. You can create, pause, and monitor webhooks using natural language instead of clicking through a dashboard. **Can I use Convoy to pause webhooks during maintenance?** Yes, you can tell your agent to pause specific endpoints instantly. This prevents your application from receiving data while you are performing updates or troubleshooting. **How do I debug failed webhooks with Convoy?** You can ask your agent to list delivery attempts for a specific event. It will pull the metadata for you, showing exactly why a delivery failed, such as a timeout or a 500 error. **Can I import my events from an API spec?** Yes, Convoy supports importing event types directly from an OpenAPI specification. This allows you to sync your entire event schema without manual entry. **Can I create multiple webhooks at once?** You can use bulk onboarding to create dozens of endpoints and subscriptions in a single request. This is much faster than creating them one by one. **Does Convoy handle event routing?** Yes, it manages the routing logic for your webhooks. You can set up specific filters and subscriptions to ensure that only the correct data reaches each endpoint. **Can I pause a specific endpoint if it's failing?** Yes! Use the `pause_endpoint` tool with the Project ID and Endpoint ID. You can later use `activate_endpoint` to resume traffic once the issue is resolved. **How do I send a webhook to multiple destinations at once?** Use the `broadcast_event` tool. It will send the event payload to all endpoints in your project that match the specified event type. **Can I see all endpoints configured in my project?** Simply run the `list_endpoints` tool with your Project ID. It will return a list of all endpoints, including their names, URLs, and current statuses.