iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
To call another agent with A2A, discover its Agent Card, check the skills, interface, capabilities, and authentication it advertises, then send a message through a compatible client. The remote agent may return a final response or, if both sides support streaming, send progress events as work proceeds. The protocol lets agents built with different frameworks communicate without exposing their internal tools or memory.
How A2A discovery and calling work
A2A is a protocol for communication between independent agents. A calling agent does not need to know how the remote agent is implemented internally. Instead, it uses the remote server’s Agent Card to learn the information needed to make a compatible request.
The interaction is a sequence: locate the card, inspect it, choose a supported interface, meet any authentication requirement, construct a message, and handle the response. Discovery is metadata exchange; it does not itself find or select the best agent for a task. Your application still needs a way to decide which agent to contact.
- Locate: Fetch the remote server’s Agent Card, commonly from
/.well-known/agent-card.json. - Inspect: Read the advertised identity, skills, capabilities, supported interfaces, and security requirements.
- Connect: Configure a client to use an interface the card declares.
- Send: Create a user-role message and submit it to the remote agent.
- Handle: Consume the final Task or Message, or process streaming events if streaming is supported and enabled.
These steps are separate checks, not a single “send and hope” operation. For example, a card can be retrievable even when the client cannot use its advertised interface. A client can support the interface while the server requires authentication the caller has not configured. And a successful connection does not mean the advertised skill accepts every possible request. Treat each stage as a boundary where your application can verify that it is still talking to the intended service in a way both sides support.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application:Upgraded attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Warranty & Support: Our Electric air duster comes with 5 years warranty and 24/7 customer service. During this period, if you have any questions,pls kindly feel free to contact us.
The released specification identifies version 1.0.0 as the latest released version. A2A has framework-independent protocol concepts, but SDK interfaces can evolve; use documentation and examples that match the SDK version you install rather than mixing APIs from older tutorials with newer ones.
What an Agent Card tells the caller
An Agent Card is metadata published by a particular A2A server. It is not a general-purpose registry or a promise that the server will accept every request. The card gives a client a basis for deciding whether and how to call that server.
A server’s card can describe its name and purpose, the skills it offers, input and output modes, supported capabilities, available interfaces, and security schemes. A2A specifies that servers must make an Agent Card available. The default well-known location is /.well-known/agent-card.json; deployments can also use a custom card path, and callers may obtain card information through a registry or preconfigured card URL or content.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Before sending work, compare the user’s need with the card’s actual skill descriptions and supported modes. If the card advertises multiple interfaces, select one it declares rather than assuming a protocol binding. Check whether the card indicates streaming support if your client needs progress updates. A card helps establish compatibility, but your application should still handle a request being rejected or a remote service being unavailable.
In practical terms, the card is the server’s published description of its contract. Its identity and description help you verify that you found the intended agent; its skills help you decide whether the agent is relevant; its modes and capabilities help you determine what kinds of input, output, or interaction to expect; and its interface declaration tells your client how to connect. The security schemes indicate whether additional steps may be required. Read these items together rather than treating a matching skill name as sufficient proof of compatibility.
The specification’s requirement is direct: “A2A Servers MUST make an Agent Card available.” That requirement makes the card a central part of discovery, but it does not turn the card into an independent directory of all agents. If your application needs to search a set of services, it still needs a registry, catalog, configured list, or another way to obtain candidate card locations. Once it has a candidate, inspect that server’s card before calling it.
Build and expose the server other agents will call
A caller can only discover and invoke an agent if the remote service exposes the protocol endpoints and a card that describes them. The official Python server tutorial composes an agent executor with a request handler and task store, then adds an Agent Card route and a JSON-RPC route to a Starlette application served by Uvicorn.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
Compose the agent and its card
Define the agent’s advertised skill and create an Agent Card with an accurate name, description, input and output modes, capabilities, and supported interface. The card should describe what the server actually accepts and does. Overstating a skill or omitting an important mode creates a mismatch: callers may send requests the implementation cannot handle.
Think of the card as part of the server’s externally visible interface, not just documentation for a human reader. If the implementation changes, update the advertised skills, modes, capabilities, interface, and endpoint so that callers are not making decisions against stale metadata. In the other direction, avoid advertising a capability merely because the underlying framework could support it; the service must actually expose and handle the corresponding interaction.
The tutorial’s example advertises JSON-RPC and uses http://127.0.0.1:9999 as its endpoint. That loopback address is for a local tutorial setup. It is reachable from the same host, not a public endpoint that another machine can discover or call. A deployed server needs an endpoint reachable by its intended clients, with its published card and route configuration aligned to that endpoint.
Wire the request handler and routes
The tutorial uses DefaultRequestHandler, an AgentExecutor, and a TaskStore. It then attaches the card and JSON-RPC routes to the Starlette application and runs the app with Uvicorn. The tutorial notes that route factories also support REST, so JSON-RPC is an example interface rather than a requirement that every A2A server use it.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
At a minimum, verify that the running application serves the card at the path you intend clients to use and that its advertised interface endpoint corresponds to a route the application actually exposes. A card can be syntactically available yet still be operationally misleading if the endpoint or supported interface does not match the server.
This separation also helps when diagnosing a server. The card route answers the discovery question—what does this server say it supports, and where is its interface?—while the protocol route is where the client sends its request. A working card route does not prove the request route is reachable, and a working request route does not help a caller that cannot locate the card. Test both from the network position your intended clients will use, rather than relying only on a local process starting successfully.
Discover the server from a Python client
The current Python client tutorial uses A2ACardResolver to fetch a card from a supplied base URL. The resolver’s default path is /.well-known/agent-card.json; when the server publishes its card elsewhere, provide the custom card path as supported by the resolver. After retrieving the card, create a client from it and configure the client for the interaction you intend to use.
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
- Choose the base URL. Use the remote server’s reachable base URL, not a loopback address copied from a local tutorial unless the client is running on that same host.
- Resolve the card. Use
A2ACardResolverwith the base URL and, if needed, the server’s custom card path. Retrieve the card before choosing an interface or constructing a request. - Check the result. Confirm that the card describes the intended agent, includes a suitable skill, and advertises an interface and capabilities your client can use.
- Configure the client. Initialize the client from the retrieved card and a
ClientConfigappropriate to the call mode.
Keep discovery separate from agent selection in your own application. If you have a known remote agent, a preconfigured card URL or content may be sufficient. If you need to choose among agents, a registry or catalog can help locate candidates, but that is an additional discovery mechanism—not a replacement for inspecting the selected server’s card.
Recommended Free Tools
A useful client-side decision order is: first confirm the candidate is the intended service; next check that its skill and modes fit the work; then choose one of its declared interfaces; finally check whether the requested interaction requires authentication or streaming support. This order avoids building a request around an assumption that the remote service never advertised. It also gives your application a clear place to stop and explain why a candidate is unsuitable, rather than making a request that is unlikely to succeed.
Send a message and handle the final response
For a basic call, the Python tutorial creates a text message with a user role using new_text_message, wraps it in a SendMessageRequest, and asynchronously iterates over client.send_message(request). In its non-streaming example, the iteration yields a final Task or Message.
- Create the user message. Use the client helper to create a text message with the user role. Keep the content focused on the task the advertised skill supports.
- Wrap it as a request. Place the message in a
SendMessageRequestfor the client call. - Call asynchronously. Iterate over the result of
client.send_message(request)using the SDK’s asynchronous flow. - Process the returned object. The non-streaming tutorial’s response is a final Task or Message. Handle the result your application receives rather than assuming every agent returns the same content or task outcome.
A Task and a Message are not interchangeable assumptions for application logic: the tutorial allows either as the final result. Design downstream handling around the response types and outcomes supported by your chosen SDK and remote agent. If your application depends on a particular output, validate that output before treating the call as successful.
Keep the boundary between request construction and result handling explicit. The message expresses the work for the remote agent; the request wrapper is what the client submits through the SDK. After the asynchronous iteration completes, your application should inspect the object it actually received and decide how it maps to its own workflow. A caller should not assume that a successful transport call automatically fulfills the user’s goal, nor should it assume that every response is plain text just because the request began as a text message.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchChoose between a single response and streaming
A single response is simpler when intermediate updates do not help the caller or user. Streaming is useful when work may take time and the interface can show progress or artifacts as they arrive. A2A supports both patterns, but streaming is capability-dependent; enabling it in the client alone does not make an unsupported server stream events.
| Consideration | Single response | Streaming |
|---|---|---|
| Client setup | Configure the client for non-streaming. | Set ClientConfig(streaming=True). |
| Response handling | Iterate to the final Task or Message. | Consume events or chunks as they arrive. |
| Useful when | Intermediate updates would not change the experience. | A long-running task or user interface benefits from progress or artifacts. |
| Requirement | The server must support the selected interface and operation. | The client configuration and server streaming capability must align. |
| Lifecycle | Follow the SDK client lifecycle. | Close the streaming client after consumption, as the tutorial does. |
The tutorial’s streaming example shows task, status, artifact, and completion events. Treat those as events to process as they arrive, not as a guarantee that every server emits every event type in that exact sequence. The remote server’s advertised capabilities and the SDK’s current interface determine what a caller can rely on.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
Choose streaming because the interaction benefits from incremental information, not simply because it is available in the SDK. A progress-aware interface can react as events arrive; a caller that only needs the completed result may prefer the single-response path. In either mode, make sure the client configuration matches the server’s declared capabilities. For streaming, also follow the SDK lifecycle and close the streaming client after consuming the stream, as shown in the tutorial. If the server does not advertise the necessary support, use the compatible non-streaming flow instead.
Meet the authentication requirements in the card
Do not assume that every agent is public or that every agent uses the same authentication scheme. Inspect the card’s security schemes and decide whether your caller can satisfy the declared requirement before sending a request.
The A2A overview gives OpenID Connect as an example: the card advertises authorization and token URLs, which a client uses in the applicable authentication flow to obtain a token. The specific scheme and URLs depend on the server. Build authentication around what that card advertises and what your identity provider requires; do not hard-code an assumption that an agent always accepts a particular credential type.
For a production caller, treat the card as an input to your connection process, not as a reason to send credentials blindly. Confirm that the endpoint and authentication configuration belong to the service you intended to call, and avoid logging tokens or other secrets while handling responses and errors. These are application security practices; the protocol’s card tells you what the server declares, not whether your application’s trust policy is sufficient.
Authentication belongs before the request is sent, because a declared requirement can determine whether the caller is ready to use that service at all. For an OpenID Connect flow, use the authorization and token URLs advertised by the card in the applicable flow; do not infer that a token obtained for a different service or scheme will be accepted. Keep the scheme-specific work separate from message construction so the client can fail early and clearly when it cannot meet the server’s declared requirements.
Common integration failures and how to diagnose them
- The card cannot be retrieved: Check the base URL, network reachability, and whether the server uses the default well-known path or a custom path. A client outside the server’s host cannot reach a card published only at a loopback address.
- The client cannot use the advertised interface: Compare the interface declared in the card with the client configuration and the routes exposed by the server. The tutorial’s JSON-RPC route is one supported example, not a universal assumption.
- The skill seems relevant but the request fails: Re-read the advertised skill and input/output modes. A card describes the server’s offering; it does not imply support for arbitrary message formats or tasks.
- Authentication is rejected: Check the security schemes declared in the card and follow the matching authentication flow. Do not retry by silently removing authentication or substituting a different scheme.
- No progress events arrive: Verify that streaming is enabled in the client and that the remote agent advertises streaming capability. Use the single-response flow when streaming is unavailable or unnecessary.
- The server works locally but not for another caller: Replace the tutorial’s loopback-only setup with a deployment endpoint reachable from the client, and ensure the card publishes the deployed endpoint.
When troubleshooting, isolate the point of failure rather than changing several pieces at once. If resolution fails, check the card URL and reachability first. If resolution succeeds but connection or request handling fails, compare the card’s declared interface with both the client configuration and the server routes. If the request reaches the agent but produces an unsuitable result, revisit the skill and input/output modes. If only streaming fails, first confirm both the client setting and the advertised server capability.
Deploying and choosing an SDK
Hosting is a deployment decision, not part of A2A’s core discovery protocol. The official tutorial index includes an ADK agent deployment tutorial for Cloud Run. It also lists sample implementations in Java, JavaScript, and C#/.NET, in addition to Python. The same high-level contract applies across those choices: publish a card, expose a supported interface, and let clients inspect the card before calling.
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
Choose a language and hosting approach based on the rest of your system, then keep the Agent Card accurate as the deployment changes. A local development endpoint, a staging service, and a public deployment are different connection targets; ensure the calling client fetches the card from the intended environment. For Python, follow the current client and server tutorials for the SDK version in use, particularly if adapting older versioned quickstarts.
The protocol flow is not tied to Cloud Run or to Python. A2A’s role is to provide a common way for independent agents to communicate; hosting, framework, and language remain implementation choices. The official materials include examples across multiple languages and a Cloud Run deployment tutorial, so teams can choose a stack that fits their service while preserving the card-and-interface contract. Do not treat a deployment tutorial as a protocol requirement or a local sample endpoint as production-ready simply because it works on the developer’s machine.
Implementation checklist
- The server makes an Agent Card available at the default well-known path or at a documented custom path.
- The card accurately describes the server’s identity, skills, modes, capabilities, interface, and security schemes.
- The advertised endpoint and interface correspond to routes the running server actually exposes.
- The caller resolves the card from a reachable URL and selects only a declared interface.
- The caller accounts for the card’s authentication requirements before sending a message.
- The application handles a final Task or Message in single-response mode, or consumes and closes a streaming client when streaming is supported.
Use this checklist as a final compatibility pass: the card should describe the deployed service, the caller should be able to reach its published endpoint, and the selected call mode should be supported by both client and server. If any one of those conditions is uncertain, resolve it before relying on the remote agent in a user-facing workflow.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteFrequently Asked Questions
How do I discover another A2A agent?
Start with a candidate server’s Agent Card. The default well-known path is /.well-known/agent-card.json, and the Python tutorial uses A2ACardResolver with a supplied base URL. A custom card path, registry, or preconfigured card URL or content may also be used. A registry can help locate candidates, but inspect the selected server’s card before calling it.
Does every A2A agent use JSON-RPC?
No. The Python server tutorial demonstrates a JSON-RPC route, and notes that route factories also support REST. Select an interface the Agent Card declares and that your client can use; do not assume every server exposes the same binding.
Can I call an A2A agent without authentication?
That depends on the server. Inspect the card’s security schemes and satisfy the declared requirement. The A2A overview gives OpenID Connect as an example, with authorization and token URLs advertised by the card; there is no single authentication assumption that fits all agents.
When should I use A2A streaming?
Use streaming when incremental progress or artifacts are useful, such as for a long-running task or a progress-aware interface. Configure the client for streaming and confirm the remote agent advertises the necessary capability. Otherwise, use the single-response flow.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Can another machine call the Python tutorial server at 127.0.0.1:9999?
Not as written when it runs on a different host. The tutorial’s 127.0.0.1 endpoint is a local loopback setup. A remote caller needs a reachable deployed endpoint, and the Agent Card must advertise that endpoint.
What does the Python client receive from a non-streaming call?
The tutorial’s non-streaming example asynchronously iterates over client.send_message(request) and yields a final Task or Message. Handle the result type your SDK and remote agent return rather than assuming every call produces the same object.
The Bottom Line
Build an A2A caller by resolving and inspecting the remote Agent Card, choosing a declared interface, satisfying its authentication requirements, and sending a user-role message through a compatible client. Use a final-response flow when intermediate updates are unnecessary, or streaming when both sides support it and progress is useful. Keep the card, server routes, and deployed endpoint aligned so callers can discover and invoke the agent reliably.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
