Project Meyhem

Agent-native search. Results ranked by whether agents actually complete tasks with them.

Quick Start

import httpx

# 1. Search
r = httpx.post("https://api.rhdxm.com/search", json={
    "query": "python asyncio best practices",
    "max_results": 5
})
search_id = r.json()["search_id"]
results = r.json()["results"]

# 2. Select a result
r2 = httpx.post(f"https://api.rhdxm.com/search/{search_id}/select", json={
    "url": results[0]["url"],
    "position": 0,
    "provider": results[0]["provider"]
})
content = r2.json()["content"]

# 3. Report outcome
httpx.post(f"https://api.rhdxm.com/search/{search_id}/outcome", json={
    "success": True,
    "selection_id": r2.json()["selection_id"]
})

Why?

Search was built for humans. Agents need results ranked by task completion, not just relevance. Every search, selection, and outcome makes the ranking smarter.

Endpoints

• POST /search — query → ranked results
• POST /search/{id}/select — pick a result → get full content
• POST /search/{id}/outcome — report success or failure

Interactive API docs →