# Treblle MCP for AI Agents AI Agent Connect > Treblle lets you send API request and response data directly to your dashboard. It's built for real-time observability, helping you see exactly what's happening with your traffic while your agent works. You get automatic masking for sensitive data and the ability to tag traces with custom metadata. ## Overview - **Category:** developer-tools - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_7VSMX7rJLXX7eBb3PrwNZUITkeSGwZgkRJYLNsyi/ai-agent-connect - **Tags:** api-monitoring, observability, api-analytics, request-logging, api-documentation ## Description Treblle gives you real-time observability on your API traffic. This Connector lets you send full request and response payloads directly to your dashboard, which helps you keep your documentation and security standards up to date. You don't have to worry about manual logging or accidentally exposing sensitive info because the system handles masking for things like passwords and credit card numbers automatically. You can also attach custom metadata like trace IDs or user IDs to help you filter through your traffic more easily. When you use this through the Vinkius catalog, you're getting a way to see exactly what's happening under the hood. It turns the guesswork of debugging into a clear view of your backend performance. ## Tools ### ingest_api_data Send full request and response payloads to Treblle. This lets you see every detail of your traffic in a dashboard while keeping sensitive data masked. ## Prompt Examples **Prompt:** ``` I just sent a request to the /login endpoint. Can you log that to Treblle? ``` **Response:** ``` I've logged the login request to Treblle. You can now view the request details, performance metrics, and documentation in your Treblle dashboard. ``` **Prompt:** ``` The user is getting a 500 error on the checkout page. Send the last request payload to Treblle. ``` **Response:** ``` The error payload for the checkout page has been sent to Treblle. It is now available in your dashboard for further investigation. ``` **Prompt:** ``` Log a successful GET request to /profile and include the user_id as a custom trace ID. ``` **Response:** ``` I've logged the GET /profile request to Treblle with the trace ID 'user_123'. The traffic data is now available for monitoring. ``` ## Capabilities ### Send full request and response payloads Your agent can push every detail of an API interaction to your dashboard for complete visibility. ### See real-time performance metrics Monitor how your API handles traffic as your agent processes requests in real-time. ### Automatically mask sensitive data The system hides passwords, credit card numbers, and SSNs before they ever hit your logs. ### Attach custom metadata Tag your ingested data with trace IDs, user IDs, or environment tags for easier filtering. ### Filter traffic history Quickly sort through your logs to find specific interactions using your custom identifiers. ## Use Cases ### Debugging a failed auth flow An agent hits an endpoint, fails, and the dev uses the Connector to see the exact 401 error payload in the dashboard. ### Monitoring high-traffic spikes A DevOps engineer watches real-time performance metrics as the agent handles a heavy load test. ### QA Testing A QA engineer captures every request during a test run to report bugs with full context and payloads. ### API Documentation A developer keeps their docs updated by letting the agent log successful interactions automatically. ## Benefits - Get instant visibility into every request your agent makes using ingest_api_data. - Keep your logs clean and secure with automatic masking of sensitive user data. - Stop guessing about errors by seeing full request and response payloads in Treblle. - Organize your traffic data easily by attaching custom trace IDs or user IDs. - Keep your API documentation accurate by automatically capturing real-world interactions. ## How It Works The bottom line is you get instant visibility into your API traffic without manual logging. 1. Connect your Treblle API Key and SDK Token in your AI client. 2. Direct your agent to send traffic data to the Connector during its tasks. 3. View the processed requests and performance metrics in your Treblle dashboard. ## Frequently Asked Questions **How does Treblle help me monitor my API traffic?** Treblle gives you a central dashboard to see every request and response your agent makes. It captures the full payload and performance data so you can see exactly what's happening. **Can Treblle mask sensitive data like passwords?** Yes, the Connector automatically masks sensitive fields like passwords, credit card numbers, and SSNs before the data is sent to your dashboard. **How do I see the performance metrics of my API calls?** Once you send your data to Treblle, you can view real-time performance metrics and response times directly in your dashboard. **Can I use Treblle to update my API documentation?** Yes, by automatically capturing real-world interactions, Treblle helps ensure your API documentation stays accurate and reflects actual traffic. **How does Treblle help with debugging AI agent errors?** It provides the full context of a failure by sending the exact request and response payloads to your dashboard, making it much easier to spot the root cause. **Can I filter my Treblle logs by custom IDs?** You can attach custom metadata like trace IDs or user IDs to your data. This allows you to filter through your logs to find specific interactions quickly. **How do I send API traffic data to my Treblle dashboard?** Use the `ingest_api_data` tool. You need to provide the JSON payload containing the server, request, and response information as defined in the Treblle schema. **Is my sensitive data like passwords or credit cards safe when ingesting?** Yes. Treblle automatically masks sensitive fields (passwords, CCs, SSNs) before the data is transmitted and stored, ensuring compliance and security. **Can I add custom identifiers like a trace-id to the ingested data?** Absolutely. You can use the optional `metadata` parameter in the `ingest_api_data` tool to include flat key-value pairs like `trace-id`, `user-id`, or `env-id`.