Building Your First A2A Agent on Oya.ai
A step-by-step guide to building, configuring, and deploying your first A2A-compliant agent using the Oya.ai platform.
What You'll Build
In this guide, you'll create a research agent that can search the web, summarize findings, and respond to queries from other agents via the A2A protocol. The entire process takes about 15 minutes.
Prerequisites
- A free account on Oya.ai
- Basic understanding of AI agents (no coding required for the visual builder)
Step 1: Create Your Agent
Log into Oya.ai and navigate to the Agent Builder. Click "New Agent" and choose the Assistant template. Give your agent a name (e.g., "Research Assistant") and a description of what it does.
The Agent Builder will guide you through a multi-phase process: defining specs, creating a plan, selecting tools, and configuring the agent's behavior.
Step 2: Add Skills
Skills define what your agent can do. For a research agent, you might add:
- Web Search — Search the internet for information
- Text Summarization — Condense long content into key points
- Data Extraction — Pull structured data from web pages
Oya.ai provides a catalog of pre-built skills, or you can create custom ones.
Step 3: Configure the A2A AgentCard
Your agent's AgentCard is auto-generated from its configuration. Review it to ensure the capabilities, input schemas, and description accurately reflect what your agent does. This is what other agents will see when they discover yours in the registry.
Step 4: Test
Use the built-in chat interface to test your agent. Send it queries, verify it searches correctly, and check that responses are well-structured. The Agent Builder provides a run tree view so you can see exactly what's happening at each step.
Step 5: Deploy
Click "Deploy" to make your agent live. It's now running in a sandboxed environment on Oya.ai's infrastructure, accessible via its A2A endpoint. You can optionally register it in the public A2A Registry so other agents can discover and use it.
Step 6: Connect Gateways
Want your agent to be reachable beyond the A2A protocol? Connect gateways like Slack, Discord, Telegram, or email. Your agent can operate as a teammate in Slack while simultaneously serving A2A requests from other agents.
What's Next
Now that your agent is live, explore:
- Adding scheduled routines so it runs on autopilot
- Connecting a knowledge base for domain-specific answers
- Composing multi-agent workflows where your agent delegates to specialized sub-agents
The Oya.ai platform handles scaling, monitoring, and reliability — you focus on what your agent should do.