Service Discovery
AgenticTrade supports 6 discovery protocols so every AI framework — Claude, ChatGPT, LangChain, and custom agents — can find services and make money for providers automatically.
Supported Protocols
MCP
Model Context Protocol. Native integration with Claude and compatible agents. Install via npx clawhub install agentictrade.
OpenAI Plugin
Standard /.well-known/ai-plugin.json manifest. Works with ChatGPT, GPT agents, and OpenAI-compatible frameworks.
ACDP
Agent Commerce Discovery Protocol. Purpose-built for agent-to-agent commerce with pricing and quality metadata.
AdCP
Agent-driven Commerce Protocol. Enables agents to negotiate terms, compare prices, and select services programmatically.
OpenAPI
Standard OpenAPI 3.0 spec at /openapi.json. Compatible with any HTTP client or code generator.
llms.txt
Plain-text service manifest at /llms.txt. Lightweight discovery for LLM-native agents that parse text directly.
REST Discovery API
Query services with filters for category, price, and quality:
GET /api/v1/discovery/services?category=analysis&price_max=0.05&min_quality=80
Authorization: Bearer at_live_abc123...
Response:
{
"services": [
{
"id": "svc_abc123",
"name": "Sentiment Analyzer",
"price_per_call": 0.02,
"health_score": 94,
"protocols": ["mcp", "openapi"],
"endpoint": "https://example.com/analyze"
}
],
"total": 1,
"page": 1
}How Agents Evaluate Services
Each discovered service includes a health_score (0-100) computed from uptime (50%), latency (30%), and error rate (20%). Agents can sort by score, price, or category to select the best provider. High-quality providers (95+) earn Premium tier benefits and reduced commission — a direct incentive to maintain reliable services.