# GeekVault Article Extractor

> GeekVault Article Extractor is a paid API for AI agents from api.geekvault.dev, paid per call via x402, $0.005/call, status unknown (last checked 2026-10-29).

Fetches a web article by URL and returns its title and main content as clean Markdown, stripping ads and navigation clutter.

## Facts

- Endpoint: GET https://api.geekvault.dev/web-tools/v1/extract-article?utm_source=zero.xyz
- Price: $0.005/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-29
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/geekvault-article-extractor-867cf7ff
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_DqS1m54jg0qiz-rp5Bj3T

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability geekvault-article-extractor-867cf7ff
```

Example prompt: Can you grab the article at https://www.theverge.com/2024/5/1/example-article and give me the clean text as markdown — no ads or sidebars?

## When to prefer this

Choose this endpoint when you need to extract clean, readable article text from a public web URL without setting up API keys or subscriptions. It is ideal for one-off or agent-driven reads of news articles, blog posts, or editorial content where you want Markdown output suitable for downstream LLM processing. Prefer alternatives if you need full-site crawling, structured data extraction, or JavaScript-heavy single-page apps.

## Known failure modes

- URL is behind a hard paywall or login — returns empty or partial content
- URL points to a non-article page (homepage, video, PDF) — markdown may be minimal or malformed
- Invalid or unreachable URL — request fails with an error response
- Rate limiting or bot detection on the target site — extraction may fail or return incomplete content

## How this service works

Network, docs, image, web, and dev utility tools paid per call via x402. No signup, no API keys, no subscriptions — just USDC on Base.

## Output

Returns a JSON object with two fields: 'title' (the article's headline as a string) and 'markdown' (the article's main body content cleaned and formatted as Markdown, free of ads, navigation, and boilerplate).

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "url"
 ],
 "properties": {
  "url": {
   "type": "string"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "title",
  "markdown"
 ],
 "properties": {
  "title": {
   "type": "string"
  },
  "markdown": {
   "type": "string"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/geekvault-article-extractor-867cf7ff/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.geekvault.dev](https://www.zero.xyz/host/api.geekvault.dev/llms.txt)