AI agent and agentic tool listings on AWS Marketplace have specific fulfillment and integration requirements. These requirements ensure customers can securely access, authenticate, and integrate your AI-powered solution after purchase, and that your listing meets AWS Marketplace validation standards.
This article explains the key fulfillment and integration concepts you must provide for AI agent–based offerings, including API access, authentication, and optional AWS integrations.
🤖 What is fulfillment for AI agent products?
Fulfillment defines how customers access your AI agent or tool after purchase and how it integrates with their environment. For AI agent products, fulfillment is typically API‑based, where customers interact with your solution through secure, vendor‑hosted endpoints.
Your fulfillment selections determine:
- How customers onboard after purchase
- How your agent or tool is accessed
- How AWS Marketplace validates your listing
⚠️ Fulfillment choices cannot be changed after publishing, so accuracy here is critical.
🔗 Fulfillment method
Defines how customers access your product after purchase and how onboarding is handled.
Options:
- SaaS – Access the software through a web browser while the seller manages the infrastructure. Customers use pre-built interfaces and workflows without needing to configure hardware or software.
- API-based agents & tools – Access AI agents through API calls to perform specific tasks. Customers integrate these agents into their applications or workflows through direct API communication.
This selection determines the integration approach and cannot be changed after publishing.
🌐 Fulfillment URL
The fulfillment URL is the entry point customers are redirected to after purchase.
Use this URL to:
- Prompt customers to sign in or create an account
- Provide onboarding instructions
- Generate or manage API credentials
Best practices:
- Use a secure HTTPS URL
- Ensure the page is accessible immediately after purchase
- Clearly guide customers on next steps
🧠 AI agent vs. AI tool
You must specify whether your listing represents a full AI agent or an AI tool.
- AI Agent: An autonomous system that uses reasoning models to plan and execute tasks without constant human input
- AI Tool: A specialized component that supports other agents
If you select AI Tool, you’ll also indicate the tool type, such as:
- MCP Server
- Knowledge base
- Guardrail
- Other supporting component
📡 API endpoint URL
This is the public endpoint where your AI agent or tool receives requests.
Requirements:
- Must be publicly accessible
- Must use HTTPS
- Must remain stable for Marketplace customers
This endpoint is used by customers, integrations, and (optionally) AWS services to interact with your solution.
📘 API usage instructions
API usage instructions explain how customers successfully integrate and use your AI agent or tool.
Include:
- Authentication steps
- Required headers or tokens
- Example request/response flows
- Usage limits or prerequisites
- Links to detailed documentation
Clear instructions reduce onboarding friction and are required for AWS Marketplace review.
🔐 Authorization method
Specify how customers authenticate when calling your API. Supported options include:
- API keys
- OAuth 2.0
If you support OAuth 2.0, provide setup steps and token usage guidance so customers can authenticate correctly.
🧩 Amazon Bedrock AgentCore Gateway integration (optional)
You may optionally enable integration with Amazon Bedrock AgentCore Gateway.
This allows customers building on AgentCore to:
- Discover your agent or tool
- Connect securely at scale
While optional, enabling this integration is recommended if your customers are using Bedrock AgentCore.
🔄 API integration protocols (optional)
If your agent or tool supports specific interaction standards, you can declare them here.
Common protocols include:
- A2A (agent-to-agent)
- MCP (Model Context Protocol)
For each selected protocol, provide clear instructions on how it’s implemented and used.
✅ AI agent requirements confirmation
AWS Marketplace requires confirmation that AI agents:
- Use reasoning-capable large language models
- Can operate autonomously without continuous human control
You must explicitly confirm that your AI agent meets these criteria to proceed with listing.
👉 What happens next
Once Section 3 is completed:
- Labra reviews your fulfillment configuration
- AWS Marketplace validates the information during listing review
- Your agent becomes available for customer onboarding after purchase
If you’re unsure how to complete any field, the Labra team can guide you through the requirements before submission.