Quickstart

From zero to a working html scrape, then an analyse call. Use json for JSON APIs (omit actions for the whole document). Use browser when the page needs JavaScript or interaction.

1. Get an API key

   1. Sign up and confirm your email.
   2. Open API keys and click Generate key.
   3. Copy the plaintext sf_… value shown on creation.

   Caution: The key is only displayed once — it is hashed at rest and cannot be recovered. If you lose it, revoke and generate a new one.

   You can also call the API from the dashboard’s built-in Run scrape page (no key required — the browser session is bound to your account), but for anything outside the browser you need a bearer token.

2. First scrape — html

   The action DSL maps each output key to a selector string: "selector@extractor" (CSS by default — no css= prefix needed).

   curl -sS -X POST https://api.scrapesilo.com/scrape/html \
     -H "Authorization: Bearer sf_…" \
     -H "Content-Type: application/json" \
     -d '{
            "url": "https://scrapesilo.com/fixtures/article",
            "actions": {
              "title": "h1",
              "body": "p",
              "cta": "a.sf-cta@href"
            }
         }'

   A completed /scrape/html HTTP 200 response returns an array (one item per distinct URL):

   [
     {
        "url": "https://scrapesilo.com/fixtures/article",
        "data": {
          "title": "How to pin an article field",
          "body": "This lead is the body field. Select the first paragraph after the heading.",
          "cta": "https://scrapesilo.com/docs/quickstart"
        },
       "tookMs": 312,
       "jobId": "job_…",
       "antibot": []
     }
   ]

   HTTP 202 instead returns job rows when the wait deadline expires or ?wait=false is used. Poll those IDs without resubmitting.

   Open the /jobs page in the dashboard to inspect the full request/result/error JSON for the run.

3. Interpretation — analyse

   For outlines, summaries, or “is X present?” use POST /analyse (or the MCP analyse tool). That is a separate surface — scrape deterministically extracts fields or a JSON document. Do not send query on /scrape.

   curl -sS -X POST https://api.scrapesilo.com/analyse \
     -H "Authorization: Bearer sf_…" \
     -H "Content-Type: application/json" \
     -d '{
            "url": "https://scrapesilo.com/fixtures/prose",
            "query": "Outline the article sections and state the main claim in one sentence."
         }'

4. Need JS or clicks? Use browser

   html is a plain HTTP fetch — no JavaScript. Switch to /scrape/browser when the page is a SPA, needs clicks/fills, or fields come back empty. Same actions tree; 5 credits per URL. See Engines.

Every scrape spends credits from your plan’s monthly pool (the free tier starts with 1,000). Cost depends on the engine — html and json are 1 credit, analyse is 3, browser is 5. Query-driven /scrape/ai is 50 credits per job.

Read a page as Markdown with MCP

Call scrape_markdown with a URL; no CSS selectors or actions are needed:

{ "url": "https://scrapesilo.com/fixtures/article", "mode": "full", "engine": "html" }

mode is full (default) or article; engine is html (default) or browser. Full mode preserves the page body; article mode uses main-article detection and can omit content. Choose browser for JavaScript-rendered pages. Read the completed result’s data.markdown. See MCP for setup, batch limits and polling.

Copy for an agent

ScrapeSilo API: https://api.scrapesilo.com
Auth: Authorization: Bearer $SCRAPESILO_API_KEY

Rules:
- MCP page reading: scrape_markdown({ url, mode: "full", engine: "html" }); no actions or selectors. Use mode="article" for main-article detection or engine="browser" for JavaScript. Defaults are full/html.
- POST /scrape/html, /scrape/json and /scrape/browser extract exact fields. Omit engine; actions required for html/browser, optional flat paths for json. Legacy POST /scrape remains supported with engine. No query field. CSS selectors, no css= prefix (e.g. h1@text). HTTP 200 returns an array of { url, data, tookMs, jobId, antibot }.
- POST /analyse interprets a page (outline, summary, flags). Required: url, query. Not a CSS plan.
- Credits per URL: html/json 1, analyse 3, browser 5. Use browser only when the page needs JS or interaction.

curl -sS -X POST https://api.scrapesilo.com/scrape/html \
  -H "Authorization: Bearer $SCRAPESILO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "url": "https://scrapesilo.com/fixtures/article",
        "actions": { "title": "h1", "body": "p", "cta": "a.sf-cta@href" }
      }'

curl -sS -X POST https://api.scrapesilo.com/analyse \
  -H "Authorization: Bearer $SCRAPESILO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "url": "https://scrapesilo.com/fixtures/prose",
        "query": "Outline the article sections and state the main claim in one sentence."
      }'

Where to next

• Engines — compare HTML, JSON, Browser and AI.
• Actions DSL — full selector + attribute reference.
• Pricing & quotas — credit costs, plans, and limits.
• MCP — drive scrapes from Claude Code or another MCP client.

Let AI select the extraction tools

Call POST /scrape/ai with { "url": "https://example.com/", "query": "Extract the page heading as title" }. AI uses Jev for rendering selection, tests actions through existing engines and returns the successful engine result with its recipe, within one 50-credit job. Use ?wait=false and poll the returned job ID for longer runs. See AI requests and limits.