# panevin-x402-api v1.2.0
> x402 payment-gated REST API for web content extraction, transformation, and AI-powered analysis on Base mainnet (eip155:8453).

## Authentication
This API uses the x402 micropayment protocol. Include an x402 payment header with each request.
Payment token: USDC on Base mainnet (eip155:8453).
Payee wallet: 0x9Bdf7621FE1DA6421E69DB0d98bFA5E30a209cf7

## Endpoints

### POST /extract/text — $0.001 USDC
Extract clean text from a URL.

Request body (JSON):
  url: string  — URL of the web page to extract text from

Response (JSON):
  url: string       — the original URL
  length: number    — character count of extracted text
  text: string      — up to 5000 chars of clean text

Example:
  POST https://api.panevin.net/extract/text
  {"url": "https://example.com"}

### POST /extract/summary — $0.003 USDC
Fetch a URL and return a brief text summary.

Request body (JSON):
  url: string  — URL of the web page to summarize

Response (JSON):
  url: string         — the original URL
  summary: string     — first 500 chars of extracted text as summary
  fullLength: number  — total character count of the page

Example:
  POST https://api.panevin.net/extract/summary
  {"url": "https://example.com"}

### POST /extract/links — $0.001 USDC
Extract all hyperlinks from a URL.

Request body (JSON):
  url: string  — URL of the web page to extract links from

Response (JSON):
  url: string        — the original URL
  links: string[]    — unique list of absolute URLs found on the page
  count: number      — number of unique links

Example:
  POST https://api.panevin.net/extract/links
  {"url": "https://example.com"}

### POST /extract/metadata — $0.001 USDC
Extract meta tags from a URL.

Request body (JSON):
  url: string  — URL of the web page to extract metadata from

Response (JSON):
  url: string         — the original URL
  title: string|null  — page <title>
  description: string|null  — meta description
  ogTitle: string|null      — og:title
  ogImage: string|null      — og:image
  keywords: string|null     — meta keywords

Example:
  POST https://api.panevin.net/extract/metadata
  {"url": "https://example.com"}

### POST /transform/markdown — $0.002 USDC
Convert an HTML page to Markdown format.

Request body (JSON):
  url: string  — URL of the web page to convert

Response (JSON):
  url: string      — the original URL
  markdown: string — Markdown representation of the page
  length: number   — character count of the Markdown

Example:
  POST https://api.panevin.net/transform/markdown
  {"url": "https://example.com"}

### POST /ai/summarize — $0.005 USDC
Fetch a URL and return an AI-generated summary using OpenRouter LLM.

Request body (JSON):
  url: string          — URL of the web page to summarize
  language: string     — (optional) language for the summary (default: English)
  maxLength: number    — (optional) max summary length in words (default: 300)

Response (JSON):
  url: string      — the original URL
  summary: string  — AI-generated summary
  language: string — language used for summary
  model: string    — LLM model used

Example:
  POST https://api.panevin.net/ai/summarize
  {"url": "https://example.com", "language": "Russian", "maxLength": 150}

### POST /ai/translate — $0.005 USDC
Fetch URL content and translate it to the target language via LLM.

Request body (JSON):
  url: string            — URL of the web page to translate
  targetLanguage: string — target language (e.g. Russian, Spanish, French)
  sourceLanguage: string — (optional) source language, auto-detected if omitted

Response (JSON):
  url: string            — the original URL
  translation: string    — translated text
  sourceLanguage: string — source language used
  targetLanguage: string — target language used
  model: string          — LLM model used

Example:
  POST https://api.panevin.net/ai/translate
  {"url": "https://example.com", "targetLanguage": "Russian"}

### POST /ai/extract-structured — $0.008 USDC
Extract structured data from a URL according to a user-defined schema via LLM.

Request body (JSON):
  url: string    — URL of the web page to extract data from
  schema: object — schema definition: {fields: [{name, type, description}]}

Response (JSON):
  url: string    — the original URL
  data: object   — extracted data matching the schema
  model: string  — LLM model used

Example:
  POST https://api.panevin.net/ai/extract-structured
  {"url": "https://example.com", "schema": {"fields": [{"name": "title", "type": "string", "description": "Page title"}]}}

### POST /ai/seo-audit — $0.10 USDC
Comprehensive SEO audit of a URL via LLM. Analyzes title, meta description, heading structure (h1-h6), image alt tags, internal/external link counts, word count, keyword density, and produces an SEO score with improvement recommendations.

Request body (JSON):
  url: string — URL of the web page to audit

Response (JSON):
  url: string             — the original URL
  score: number           — SEO score (1-100)
  issues: string[]        — list of detected SEO issues
  recommendations: string[] — 3 actionable improvement recommendations
  metrics: object         — raw metrics (headings, images, links, keywords, etc.)
  model: string           — LLM model used

Example:
  POST https://api.panevin.net/ai/seo-audit
  {"url": "https://example.com"}

### POST /ai/compare — $0.15 USDC
Compare two web pages via LLM. Evaluates topic similarity, content depth, readability, and key differences; declares a winner.

Request body (JSON):
  url1: string — first URL to compare
  url2: string — second URL to compare

Response (JSON):
  url1: string                   — first URL
  url2: string                   — second URL
  comparison: object             — {similarity: float, differences: string[], winner: string, analysis: string}
  model: string                  — LLM model used

Example:
  POST https://api.panevin.net/ai/compare
  {"url1": "https://example.com", "url2": "https://example.org"}

### POST /ai/sentiment — $0.05 USDC
Analyze the sentiment of any URL content via LLM.

Request body (JSON):
  url: string      — URL of the web page to analyze
  language: string — (optional) expected content language (default: en)

Response (JSON):
  url: string        — the original URL
  sentiment: string  — overall sentiment: positive | negative | neutral
  confidence: number — confidence score (0.0-1.0)
  phrases: object[]  — key phrases with individual sentiment labels
  tone: string       — emotional tone description (e.g. "informative", "enthusiastic")
  model: string      — LLM model used

Example:
  POST https://api.panevin.net/ai/sentiment
  {"url": "https://example.com", "language": "en"}

## Discovery
- x402 manifest: https://api.panevin.net/.well-known/x402
- Health check:  https://api.panevin.net/health