A2A (Agent2Agent) is an open protocol for communication between Agents. Through A2A, different Agents can discover each other's capabilities, send messages, and return execution results.A2A Debugger is a visual tool for debugging A2A Agents. You can connect to an A2A Agent, send test messages, view the Agent's returned content, execution status, and debug logs, helping developers quickly troubleshoot issues with Agent integration, message delivery, and returned results.A2A Debugger is suitable for the following scenarios:Check whether an A2A Agent can connect normally
View the Agent's name, description, capabilities, and Skills
Send text or attachments to the Agent to verify its response
Troubleshoot connection or request issues such as authentication, Header, file types, etc.
View request and response details to identify the cause of abnormal returned content
We recommend using the latest Apidog client to experience the full features of AI Agent Debugger.
Connecting to an A2A Agent#
Enter the Agent Card URL at the top of the page, then click Connect.After a successful connection, the status at the top will display as Connected. At this point, you can view Agent information and start sending messages.If you need to switch to another Agent, you can click Disconnect and then re-enter the Agent Card URL.Writing Test Messages#
Write test input to be sent to the Agent in the Messages tab.The message editing area is divided into left and right sections:Left side: Enter message text, upload attachments, configure Metadata
Right side: View the corresponding raw request content
Sending Messages#
After completing the message configuration, click Send.If no Agent is currently connected, A2A Debugger will first attempt to connect; it will send the message after the connection is successful.After sending, you can see the user message and the Agent's returned content in the message area.Viewing Reply Messages#
The message area displays user messages and Agent replies in the current session.The content returned by the Agent may be plain text, or it may include task status, execution results, or generated artifacts. You can quickly determine whether the Agent correctly understood and processed the request through the message display area.Click any debug message to view details on the right. The detail panel supports three views:Preview: Display main content in a more readable way
Content: View text or core data in the message
Raw Data: View the complete raw JSON
What is the distance between the Earth and the Moon?
Viewing the Agent Card#
The Agent Card tab in the debugging panel is used to view the capability description of the currently connected Agent.Agent name and description
Agent address, protocol version, version number
Agent-supported capabilities, such as whether it supports streaming responses, push notifications, status history, etc.
Default input types and output types
Skills declared by the Agent
Agent Card validation errors
Adding Attachments#
If the current Agent supports file input, you can click the attachment button below the message input area to upload files.A2A Debugger determines whether uploading attachments is allowed based on the input types declared by the Agent:If the Agent only supports text input, the attachment button is unavailable
If the Agent only supports specific file types, you can only upload files of matching types
If the Agent supports any file type, you can upload any file
Configuring Authentication#
If the A2A Agent requires authentication, you can configure the authentication method in the Authentication tab.The following authentication types are supported:| Authentication Method | Description |
|---|
| None | No authentication information added |
| Bearer Token | Authenticate using a token |
| Basic Auth | Authenticate using username and password |
| API Key | Pass API Key via a specified Header |
If an unauthorized error is returned when sending a message, please check whether the authentication method, token, username/password, or API Key is correct.In the Header tab, you can add custom request Headers.Custom Headers are commonly used in the following scenarios:Passing business parameters such as tenant ID, environment identifier, etc.
Authentication through gateways, proxies, or internal services
Debugging Agent services that require special request headers
FAQ#
What should I do if the connection fails?#
Please check the following:1.
Whether the Agent Card URL is correct and accessible from the current environment.
2.
Whether the Agent service is running.
3.
If authentication is required, whether the correct authentication information has been configured in the Authentication tab.
4.
Whether additional Headers are needed, such as tenant ID, environment identifier, or gateway authentication information.
5.
Whether the content returned by the Agent Card is valid JSON.
Unable to upload attachments after a successful connection?#
This is usually because the current Agent has not declared support for file input, or only allows uploading specific file types.Please view the input types supported by the Agent in the Agent Card tab, and confirm whether the uploaded file type matches.How do I view the complete request and response?#
Click the corresponding request or response log in the Messages tab of the debugging panel, then switch to the Raw Data view to view the complete JSON.What should I do if the content returned by the Agent displays abnormally?#
You can troubleshoot in the following order:1.
Check the preview content in the debug message to confirm whether the Agent has returned a result.
2.
Switch to the Raw Data view to view the complete response.
3.
Check whether the input types, output types, and Skills declared in the Agent Card meet expectations.
4.
Check for validation error prompts in the message or debug details.
