deploy clean-read x402 endpoint

Dockerfile

@@ -0,0 +1,17 @@
+# HuggingFace Space (Docker SDK) for the clean-read x402 endpoint.
+# Free, no credit card, stable URL. Sleeps after ~48h idle, wakes on request.
+FROM python:3.11-slim
+
+WORKDIR /app
+ENV PYTHONUNBUFFERED=1 PORT=7860
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY scripts ./scripts
+
+EXPOSE 7860
+# --proxy-headers + trust all forwarders so the x402 `resource` URL is built as
+# https:// (HF terminates TLS at its proxy and forwards plain http internally);
+# a scheme mismatch can make strict x402 verifiers reject the payment.
+CMD ["uvicorn", "scripts.clean_read_api.main:app", "--host", "0.0.0.0", "--port", "7860", "--proxy-headers", "--forwarded-allow-ips", "*"]

README.md

@@ -0,0 +1,23 @@
+---
+title: clean-read
+emoji: 📖
+colorFrom: green
+colorTo: blue
+sdk: docker
+app_port: 7860
+pinned: false
+license: mit
+---
+
+# clean-read API
+
+URL → clean Markdown for AI agents. Fetches any page, strips nav/ads/boilerplate
+with trafilatura, returns the main content as Markdown with title + word count.
+
+Paid via **x402** micropayments (USDC on Base, Dexter facilitator, 0% fee):
+
+- `GET /` — service info (free)
+- `GET /health` — health check (free)
+- `POST /read` — `{"url": "..."}` → Markdown — `$0.005` USDC
+
+Unpaid requests get `HTTP 402` with payment requirements. Source: github.com/eltociear/my-molt-agent

requirements.txt

@@ -0,0 +1,6 @@
+# clean-read x402 API (x402 protocol v2, official SDK).
+fastapi>=0.115.0
+uvicorn[standard]>=0.23.0
+httpx>=0.28.1
+trafilatura>=2.0.0
+x402[fastapi,evm,extensions]>=2.12.0

scripts/clean_read_api/main.py

@@ -0,0 +1,173 @@
+#!/usr/bin/env python3
+"""clean-read x402 API — URL → clean Markdown extraction as a paid service (x402 v2).
+
+Deploy: uvicorn scripts.clean_read_api.main:app --host 0.0.0.0 --port $PORT
+Local:  uvicorn scripts.clean_read_api.main:app --port 8403
+
+Endpoints:
+  GET  /        — Service info (free)
+  GET  /health  — Health check (free)
+  POST /read    — Fetch URL, strip boilerplate, return main content as Markdown. $0.005/call (x402)
+
+Same stack as scripts/x402_api/main.py (skill-audit): official x402 v2 SDK,
+USDC on Base mainnet, Dexter facilitator (zero-gate, 0% seller fee, auto-lists
+on Bazaar discovery after first settled payment).
+
+Why this service: agents constantly need "give me the readable text of this page"
+(the Jina-reader use case) — extraction via trafilatura, priced at high-volume
+$0.005 like tokenguard (our best-selling price point).
+"""
+import os
+from datetime import datetime
+from fastapi import FastAPI, HTTPException
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+from typing import Optional
+
+WALLET = os.environ.get("BASE_WALLET_ADDRESS", "0x5bCDA55247B238a573A968B234F788a2D35664Dd")
+BASE_MAINNET = "eip155:8453"
+FACILITATOR_URL = os.environ.get("FACILITATOR_URL", "https://x402.dexter.cash")
+
+app = FastAPI(
+    title="clean-read API",
+    description="URL to clean Markdown for AI agents. x402 v2 micropayments on Base.",
+    version="1.0.0",
+)
+app.add_middleware(CORSMiddleware, allow_origins=["*"], allow_methods=["*"], allow_headers=["*"])
+
+_x402_available = False
+try:
+    from x402.http import FacilitatorConfig, HTTPFacilitatorClient, PaymentOption
+    from x402.http.middleware.fastapi import PaymentMiddlewareASGI
+    from x402.http.types import RouteConfig
+    from x402.mechanisms.evm.exact import ExactEvmServerScheme
+    from x402.server import x402ResourceServer
+    from x402.extensions.bazaar import declare_discovery_extension, OutputConfig
+
+    facilitator = HTTPFacilitatorClient(FacilitatorConfig(url=FACILITATOR_URL))
+    server = x402ResourceServer(facilitator)
+    server.register(BASE_MAINNET, ExactEvmServerScheme())
+
+    ext = declare_discovery_extension(
+        input={"url": "https://example.com/article"},
+        input_schema={
+            "properties": {
+                "url": {"type": "string", "format": "uri", "description": "Page to fetch and clean"},
+                "include_links": {"type": "boolean", "description": "Keep hyperlinks in the Markdown (default true)"},
+            },
+            "required": ["url"],
+        },
+        body_type="json",
+        output=OutputConfig(example={
+            "url": "https://example.com/article",
+            "title": "Article title",
+            "markdown": "# Article title\n\nMain content…",
+            "word_count": 1234,
+        }),
+    )
+    ext["bazaar"]["info"]["input"]["method"] = "POST"
+
+    routes = {
+        "POST /read": RouteConfig(
+            accepts=[PaymentOption(scheme="exact", pay_to=WALLET, price="$0.005", network=BASE_MAINNET)],
+            mime_type="application/json",
+            description="Fetch a URL and return its main content as clean Markdown (boilerplate stripped)",
+            extensions=ext,
+        ),
+    }
+    app.add_middleware(PaymentMiddlewareASGI, routes=routes, server=server)
+    _x402_available = True
+except Exception as e:  # pragma: no cover
+    print(f"  x402 v2 init warning: {type(e).__name__}: {e}")
+
+
+class ReadRequest(BaseModel):
+    url: str
+    include_links: Optional[bool] = True
+    max_size: Optional[int] = 2_000_000  # 2MB raw HTML cap
+
+
+@app.get("/")
+async def root():
+    return {
+        "service": "clean-read API",
+        "version": "1.0.0",
+        "description": "URL → clean Markdown. Boilerplate/nav/ads stripped, main content only. Built for AI agents.",
+        "endpoints": {
+            "GET /": "Service info (free)",
+            "GET /health": "Health check (free)",
+            "POST /read": "Fetch URL → Markdown ($0.005 USDC)",
+        },
+        "payment": {
+            "method": "x402",
+            "x402_version": 2,
+            "currency": "USDC",
+            "network": "Base (eip155:8453)",
+            "facilitator": FACILITATOR_URL,
+            "wallet": WALLET,
+            "x402_enabled": _x402_available,
+        },
+    }
+
+
+@app.get("/health")
+async def health():
+    return {"status": "ok", "timestamp": datetime.utcnow().isoformat() + "Z", "x402_enabled": _x402_available}
+
+
+@app.post("/read")
+async def read_url(req: ReadRequest):
+    url = req.url
+    if not url or not url.startswith(("http://", "https://")):
+        raise HTTPException(400, "valid http/https URL required")
+
+    import httpx
+    try:
+        async with httpx.AsyncClient(follow_redirects=True, timeout=20.0) as client:
+            resp = await client.get(url, headers={"User-Agent": "Mozilla/5.0 (compatible; clean-read/1.0; +https://eltociear-clean-read.hf.space)"})
+            resp.raise_for_status()
+    except httpx.HTTPStatusError as e:
+        raise HTTPException(502, f"upstream returned {e.response.status_code}")
+    except Exception as e:
+        raise HTTPException(502, f"fetch failed: {type(e).__name__}: {e}")
+
+    html = resp.text
+    if len(html) > req.max_size:
+        html = html[: req.max_size]
+
+    import trafilatura
+    markdown = trafilatura.extract(
+        html,
+        output_format="markdown",
+        include_links=bool(req.include_links),
+        include_tables=True,
+        favor_recall=True,
+    )
+    if not markdown:
+        raise HTTPException(422, "could not extract main content from this page")
+
+    title = None
+    try:
+        meta = trafilatura.extract_metadata(html)
+        if meta:
+            title = meta.title
+    except Exception:
+        pass
+
+    return {
+        "url": str(resp.url),
+        "title": title,
+        "markdown": markdown,
+        "word_count": len(markdown.split()),
+        "fetched_at": datetime.utcnow().isoformat() + "Z",
+    }
+
+
+if __name__ == "__main__":
+    import uvicorn
+    port = int(os.environ.get("PORT", 8403))
+    print(f"\n  clean-read API (x402 v2) starting on :{port}")
+    print(f"  x402: {'ENABLED' if _x402_available else 'DISABLED (pip install x402[fastapi,evm,extensions])'}")
+    print(f"  Facilitator: {FACILITATOR_URL}")
+    print(f"  Wallet: {WALLET}\n")
+    uvicorn.run(app, host="0.0.0.0", port=port)