Website Text Extractor — Clean Text & Source Fingerprints

Turn web pages into clean text for AI/RAG, content migration and ETL. Get title, author/date tags, word count and a SHA-256 content fingerprint in one JSON. Batch up to 100 URLs, up to 50,000 text characters per page. Launch price: $0.50/1,000 URL reports.

Get readable static page text and a stable SHA-256 fingerprint. Use the text for RAG or the fingerprint to decide whether content needs re-indexing.

When to use it

• Load explicitly selected public pages into a RAG pipeline.
• Export page text for content migration.
• Compare text fingerprints across recurring runs.

Start with this input

{
  "urls": [
    "https://example.com",
    "https://www.iana.org/help/example-domains"
  ]
}

Click Start, then read the Output table or download JSON/CSV. Single and batch targets are supported. Duplicate targets are removed before network work.

What you get

Clean text, title, word count, explicit author/dates, text fingerprint, extraction method, HTTP evidence and warnings.

One record per unique processed target, including observedAt, status and warnings. Missing data, unsupported sources, lookup failures, timeouts and invalid input are distinct states. Partial results remain usable. Dataset views show focused fields; the complete JSON record remains available through the dataset API.

{
  "schemaVersion": 1,
  "coreVersion": "0.1.0",
  "product": "clean-website-text",
  "normalizedInput": "http://example.com/",
  "observedAt": "2026-09-18T19:57:04.215Z",
  "status": "success",
  "warnings": [],
  "metadata": {
    "title": {
      "status": "found",
      "value": "Example Domain"
    },
    "description": {
      "status": "not_found"
    },
    "canonical": {
      "status": "not_found"
    },
    "author": {
      "status": "not_found"
    },
    "publishedDate": {
      "status": "not_found"
    },
    "modifiedDate": {
      "status": "not_found"
    },
    "wordCount": {
      "status": "found",
      "value": 19
    },
    "textSha256": "8c1e8564424fdb68b8b7bdff3e16173a2e3599e9b71620637251486c5c4d5ed6"
  }
}

Price and spending control

$0.0005 per URL report ($0.50/1,000), plus $0.0005 per run at the default 256 MB. Platform usage is included. No separate per-page or automatic dataset-item charge.

Targets	Maximum event price
1	$0.001000
10	$0.005500
100	$0.050500

Valid attempted diagnostics are charged, including HTTP errors, missing fields and DNS/TLS/timeout failures. Invalid input, whole-target private-address rejection and targets not started at the run-time cap receive no result event; a run-start event can still apply. A started inspection reaching a body/request/time cap is a charged diagnostic.

Set Maximum cost per run in Console or maxTotalChargeUsd in the API. The runtime reserves affordable targets before lookup and records omitted targets in METRICS. maxItems adds an independent item cap. Actor-start is per GB of memory with a minimum of one event; larger memory can increase it.

Call from your workflow

JavaScript:

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('gertner-data/clean-website-text').call(
  {"urls":["https://example.com","https://www.iana.org/help/example-domains"]}, { maxTotalChargeUsd: 0.001500 }
);
const { items } = await client.dataset(run.defaultDatasetId).listItems();

Python:

import os
from apify_client import ApifyClient
client = ApifyClient(os.environ['APIFY_TOKEN'])
run = client.actor('gertner-data/clean-website-text').call(run_input={"urls":["https://example.com","https://www.iana.org/help/example-domains"]}, max_total_charge_usd=0.02)
rows = list(client.dataset(run['defaultDatasetId']).iterate_items())

HTTP: POST JSON input to https://api.apify.com/v2/acts/gertner-data~clean-website-text/runs?maxTotalChargeUsd=0.02 with your bearer token in the Authorization header. Wait for completion and GET /v2/datasets/{defaultDatasetId}/items. Use asynchronous runs for larger batches.

For recurring checks, save your own task and attach an Apify schedule. For downstream integrations, consume defaultDatasetId after RUN.SUCCEEDED and route each item's structured status.

API coverage and run settings

Up to 50,000 plain-text characters per selected URL. Main/article elements take priority, with body fallback. Deterministic boilerplate filtering and a SHA-256 fingerprint support repeatable indexing workflows.

Public HTTP(S) on ports 80/443; private/reserved addresses and unsafe redirects are blocked. Processes public static content over HTTP(S). Per-page, aggregate-body, request and total-time limits keep costs bounded; limit codes remain visible. UTF-8 expected. Maximum 100 unique targets per run. Inspect only sources you are authorized to access.

Support: Evgeny Gertner.