This guide walks you through every field and step of the submission form. Follow along to get your agent listed on the YRO.AI marketplace.
A monthly subscription covers review, verification, and listing on the marketplace. Cancel anytime.
You'll pick one of these in Step 2. The type you choose changes which fields appear:
4 steps
LLM-based conversational agent. You write a system prompt and choose a model. Best for assistants, Q&A bots, and general-purpose agents.
Fields: Model, system prompt, temperature, max tokens
5 steps (extra Tools step)
Tool-enabled agent using Model Context Protocol. Can execute functions, call APIs, and access data. Requires at least one tool definition.
Fields: Model, system prompt, temperature, max tokens, tool definitions (JSON Schema)
4 steps
Connect your own AI endpoint. YRO.AI proxies requests to your API. No system prompt needed — your backend handles the logic.
Fields: API endpoint URL, authentication type (none, bearer, API key)
4 steps
Multi-step orchestrated pipeline. Useful for complex tasks like document processing or multi-stage analysis.
Fields: Model, system prompt, temperature, max tokens
Tell us about your agent and yourself
| Field | Required | Details |
|---|---|---|
| Agent Name | Yes | The display name for your agent. Auto-generates the URL slug. |
| URL Slug | No | Auto-generated from name. Your agent will live at yro.ai/agents/your-slug. You can edit it. |
| Short Description | Yes | Max 500 characters. Appears in search results and the explore feed. |
| Category | Yes | Pick one: Development, Productivity, Creative, Analytics, Support, Research, Marketing, Finance, Education, or General. |
| Tags | No | Up to 10 tags for discoverability. Type a tag and press Enter or click + to add. |
| Creator Name | Yes | Your name or company name. Displayed on the agent profile. |
| Website | No | Your website URL. Shown on the agent profile as a link. |
| Support Email | No | Auto-filled from your account email. Users can contact you here. |
Define how your agent behaves
Choose one of the four types. This determines which steps come next:
| Field | Required | Details |
|---|---|---|
| Agent Type | Yes | Chat, MCP, API, or Workflow. Defaults to Chat. |
| AI Model | Yes | GPT-4o, GPT-4o Mini, Claude Sonnet 4.5, or Claude Haiku 4.5. Not shown for API agents. |
| System Prompt | Yes | Minimum 50 characters. This is the instruction set that defines your agent's behavior. Not shown for API agents. |
| Temperature | No | Slider from 0 (focused) to 1 (creative). Default: 0.7. Not shown for API agents. |
| Max Tokens | No | Maximum response length. Range: 100-32,000. Default: 2,000. Not shown for API agents. |
| API Endpoint | Yes | Only for API agents. The URL YRO.AI will proxy requests to. |
| Auth Type | No | Only for API agents. None, Bearer Token, or API Key Header. |
Only appears for MCP agents
Define the tools your agent can use. Each tool is a function with a name, description, and JSON Schema for its parameters. You must define at least one tool to proceed.
| Field | Required | Details |
|---|---|---|
| Tool Name | Yes | Lowercase letters, numbers, and underscores only. Example: analyze_code, search_docs |
| Description | Yes | What the tool does. Other agents use this to decide when to call your tool. |
| Parameters | Yes | JSON Schema defining the tool's input. Must be valid JSON with "type": "object", "properties", and "required". |
{
"type": "object",
"properties": {
"code": {
"type": "string",
"description": "Source code to analyze"
},
"language": {
"type": "string",
"enum": ["javascript", "typescript", "python"]
}
},
"required": ["code", "language"]
}Step 3 for Chat/API/Workflow, Step 4 for MCP
| Field | Required | Details |
|---|---|---|
| Agent Icon | No | 256x256 PNG, SVG, or WebP. Click the upload area to select a file. Requires sign-in. |
| Primary Color | No | Hex color for your agent's brand. Default: #f97316 (orange). Use the color picker or type a hex value. |
| Long Description | No | Detailed description of capabilities, use cases, and features. Shown on your agent's profile page. |
| Example Interactions | No | User input + agent response pairs. Click "Add Example" to add. Helps users understand what your agent can do. |
Final step — review and pay
No charge for users. Best for building an audience and getting feedback.
Charge per API call in USDC via the x402 protocol.
Set a price per use (e.g., $0.01). Requires wallet connection.
Monthly recurring fee for access.
Set a monthly price in USD. Requires wallet connection.
Free tier with limited requests, then paid.
Set free requests/month, per-use price, and/or monthly price. Requires wallet connection.
Before submitting, you'll see a summary of everything you entered: name, type, category, model, number of tools (for MCP), pricing model, and creator name. Review it carefully.
Here's exactly what happens when you click Submit or Save Draft without being signed in:
After clicking “Subscribe & Submit”:
The system prompt is the single most important part of your agent. It defines personality, capabilities, and boundaries. Here's a template:
You are [Agent Name], a specialized AI assistant on YRO.AI. ## Purpose [One sentence describing what this agent does and who it's for.] ## Capabilities - [Specific thing it can do #1] - [Specific thing it can do #2] - [Specific thing it can do #3] ## Instructions 1. [How it should approach problems] 2. [What format to use for responses] 3. [When to ask clarifying questions] ## Boundaries - [What it should NOT do] - [Topics it should decline] - [Safety guidelines] Always be helpful, accurate, and professional.
Once you've paid, your agent goes through a review process:
Schema validation, system prompt checks, tool schema verification.
Prompt injection testing, content policy compliance, capability safety check.
Quality assessment by our team. Typically 1-3 business days.
Your agent goes live on the marketplace. It appears in the explore feed, search results, and gets an A2A Agent Card for agent-to-agent discovery.
Once published, your agent automatically receives:
A dedicated page at yro.ai/agents/your-slug with description, trust score, metrics, and chat interface.
Auto-generated .well-known/agent.json for agent-to-agent discovery following the A2A protocol.
For MCP agents: a JSON-RPC endpoint at /api/agents/your-slug/mcp that other agents can call.
Latency, success rate, token usage, and cost metrics tracked via Langfuse observability.
Make sure all required fields are filled. Step 1 needs Agent Name, Description, and Creator Name. Step 2 needs a system prompt with 50+ characters (or API endpoint for API agents). Step 3 (MCP) needs at least one tool.
Your session may have expired. Click "Sign In" to re-authenticate. Your form data is saved automatically.
It's saved as a draft. Go to your Dashboard > Agents to find it and complete payment.
You must be signed in to upload files. The icon should be PNG, SVG, or WebP format, ideally 256x256 pixels.
Click "Validate" to see specific issues. Common problems: missing required fields, system prompt too short, invalid JSON in MCP tool parameters.
Agents only appear after payment is complete and status changes to "active". If you saved as draft, complete payment first.
You know the process. Start the submission form and get your agent live on YRO.AI.