Website Favicon Discovery

Website favicon extraction for catalogue enrichment: supply public website URLs and receive one structured row per website with favicon URLs, declared sizes, MIME types, rel values and source metadata. Use the results to add source-linked icons to website directories, CRM company records or brand inventories.

What does it do?

The Actor fetches each supplied HTML page and reads supported icon declarations. It resolves relative URLs against the page's first valid base declaration or final redirected URL. It does not download, resize, convert or validate images.

Supported declarations are icon (including shortcut icon), apple-touch-icon, apple-touch-icon-precomposed and mask-icon. Optional root favicon inference is clearly marked as a fallback. Web manifests, dynamically inserted DOM elements and generic page images are not collected.

Who is it for?

• Directory operators enriching recurring website catalogues.
• Developers mapping company domains to declared icon assets.
• Brand researchers comparing published icon declarations across websites.

This is not an icon generator or a brand-logo detector.

Why use Website Favicon Discovery?

Keep the page URL, final hostname and icon attributes together rather than losing provenance in a list of images. A single row includes every retained declaration so downstream joins remain one website to one record. Exact duplicate declarations are removed, while the same URL with different rel, type or sizes is preserved.

Getting started

1. Enter public website URLs in Website URLs.
2. Choose a maximum number of websites to attempt.
3. Leave inferred fallback disabled when you only want declared icons.
4. Run the Actor and open the default dataset.
5. Export JSON to retain nested icons, or select flattened fields for a spreadsheet.

Example input:

{
  "startUrls": [{"url":"https://apify.com"},{"url":"https://github.com"}],
  "maxItems": 10,
  "includeFallback": false
}

Inputs

Field	Default	Meaning
startUrls	Required	1–1000 fully qualified public HTTP(S) URLs. Each entry is an object with a url field.
maxItems	10	Global cap on distinct normalized URLs attempted, in input order; range 1–1000. Failures count. Zero/unlimited is unsupported.
includeFallback	false	Only when no supported declaration exists, return final-origin /favicon.ico as an unverified inference.

Only GET requests are supported. URL entries accept url only; method, headers, payload and userData controls are rejected even if a client editor offers them.

Duplicate normalized URLs are processed once, ignoring fragments. Distinct paths are distinct inputs even on the same domain. Only standard HTTP(S) ports are supported. Credentials and private/reserved network destinations are rejected, including redirect destinations.

Extracted data

Field	Meaning
inputUrl	Normalized input URL
finalUrl	URL after redirects
hostname	Final page hostname
httpStatus	Successful HTML response status
pageTitle	First title, up to 1000 characters, or null
icons	Array of icon URL, rel, type, sizes, source and verified flag
iconCount	Number of retained icons
primaryIconUrl	First declaration whose rel includes icon, otherwise first retained declaration; null when none
status	found (including inferred fallback) or no_icons
truncated	True when more than 200 unique valid declarations were found
scrapedAt	UTC extraction timestamp

icons[].type and icons[].sizes are author-declared strings, not measured properties. They are null when absent. rel is lowercased and whitespace-normalized. source is html or fallback. verified is always false: a declaration does not prove that an asset exists or is usable. Non-HTTP(S) icon URLs, including embedded data URLs, are excluded.

Output example

A representative GitHub page result has this structure; timestamps vary:

{
  "inputUrl": "https://github.com/",
  "finalUrl": "https://github.com/",
  "hostname": "github.com",
  "httpStatus": 200,
  "pageTitle": "GitHub · Change is constant. GitHub keeps you ahead. · GitHub · GitHub",
  "icons": [
    {"url":"https://github.githubassets.com/assets/pinned-octocat-093da3e6fa40.svg","rel":"mask-icon","type":null,"sizes":null,"source":"html","verified":false},
    {"url":"https://github.githubassets.com/favicons/favicon.png","rel":"alternate icon","type":"image/png","sizes":null,"source":"html","verified":false},
    {"url":"https://github.githubassets.com/favicons/favicon.svg","rel":"icon","type":"image/svg+xml","sizes":null,"source":"html","verified":false}
  ],
  "iconCount": 3,
  "primaryIconUrl": "https://github.githubassets.com/favicons/favicon.png",
  "status": "found",
  "truncated": false,
  "scrapedAt": "2026-10-26T07:00:00.000Z"
}

See the dataset for the actual current declarations. The Actor does not rank icons by image quality, pixel size or browser appearance.

How much does it cost to extract website favicons?

Pricing is pay-per-event: a one-time start fee plus one item event per successfully inspected website row. This includes inspections finding no icons: the delivered result records that the inspected HTML has no supported icon declaration. The number of icons inside a row does not multiply the charge. Failed websites have no item charge. The start fee still applies after valid input is accepted, even if sources fail.

The one-time start fee is $0.0015 on every tier. Per inspected website:

Apify Store spend tier	Item price
FREE	$0.000414
BRONZE	$0.00036
SILVER	$0.0002808
GOLD	$0.000216
PLATINUM	$0.000216
DIAMOND	$0.000216

At BRONZE, 1 inspected website costs about $0.00186, 10 cost $0.0051, and 100 cost $0.0375, including one start fee. These Actor-charge estimates exclude taxes and any separately applicable Apify account charges. Higher qualifying aggregate monthly Apify Store spend tiers receive the displayed discounts; these are not private per-Actor volume discounts.

Limits and failure behavior

Each page allows at most five redirects and 10 seconds per request attempt. A transient network error, 429 or selected 5xx response can be retried once. The HTML download is capped at 2 MiB; larger pages fail instead of silently returning incomplete declarations. Pages with more than 1024 nested elements also fail rather than exhaust parser resources. Streaming parsing retains no full DOM. Up to four websites are fetched concurrently; input attempts are selected in order, but result rows can arrive in completion order.

The Actor uses direct HTTP, no browser and no paid proxy fallback. Protected, login-only, JavaScript-only or geographically restricted pages may fail. HTML responses without icons are valid no_icons results. Non-HTML responses and HTTP errors fail the source. If any website fails, the run is FAILED after the remaining selected inputs are attempted; available dataset rows are partial results, not a complete successful batch.

There are at most 200 icon declarations per website; truncated=true explicitly identifies overflow. Source changes, geo responses and dynamically generated markup can affect which declarations appear. The optional fallback may point to a missing file.

Integrations

Join results to your catalogue by inputUrl or hostname. Keep finalUrl to audit redirected domains. Store the nested icon array in your database, or use primaryIconUrl in a review queue before displaying third-party images.

Apify dataset exports and integrations can deliver results to spreadsheets, automation platforms or your own data pipeline. To compare changes, schedule runs with Apify and compare saved datasets downstream; this Actor does not keep historical state, issue alerts or perform change detection itself.

API usage

Start a run once, retain its run ID, and wait for that same run to finish before reading its default dataset. A start response is metadata, not extracted rows.

cURL

curl -X POST 'https://api.apify.com/v2/acts/automation-lab~website-favicon-discovery/runs' \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"startUrls":[{"url":"https://apify.com"}],"maxItems":1}'

JavaScript

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({token: process.env.APIFY_TOKEN});
const run = await client.actor('automation-lab/website-favicon-discovery').call({
  startUrls: [{url: 'https://apify.com'}], maxItems: 1
});
if (run.status !== 'SUCCEEDED') throw new Error(`Run ${run.id}: ${run.status}`);
const {items} = await client.dataset(run.defaultDatasetId).listItems({limit: 20});
console.log(items);

Python

import os
from apify_client import ApifyClient
client = ApifyClient(os.environ['APIFY_TOKEN'])
run = client.actor('automation-lab/website-favicon-discovery').call(
    run_input={'startUrls': [{'url': 'https://apify.com'}], 'maxItems': 1})
if run['status'] != 'SUCCEEDED':
    raise RuntimeError(f"Run {run['id']}: {run['status']}")
print(client.dataset(run['defaultDatasetId']).list_items(limit=20).items)

MCP setup

Claude Code

claude mcp add --transport http apify \
  "https://mcp.apify.com?tools=automation-lab/website-favicon-discovery"

Claude Desktop, Cursor, and VS Code

Use the equivalent HTTP MCP configuration supported by your client:

{
  "mcpServers": {
    "apify": {
      "type": "http",
      "url": "https://mcp.apify.com?tools=automation-lab/website-favicon-discovery"
    }
  }
}

Authenticate using the client's supported Apify sign-in or token configuration. Discover actual names and schemas with scoped tools/list; Actor selection also includes run and storage tools, not exactly one tool.

Example prompt: “Extract declared favicon URLs from https://apify.com and https://github.com. Show the final domains, primary icon URLs and unverified status.”

For an authorized invocation, start once and retain the returned run/storage IDs. If still running after the server wait (at most 45 seconds), poll the same run with 2, 4, 8 then at most 10 seconds backoff, stopping after a 120-second total deadline. On a client timeout, check the known run rather than starting again. At deadline, report pending with the run ID. A failed run's rows are partial.

After success, read untransformed dataset pages of 20 rows, starting at offset 0, selecting only needed fields. Stop at exhaustion, 100 source rows or 64 KiB of serialized UTF-8 results admitted to model context, whichever occurs first. Enforce the byte limit host-side; if your client cannot intercept oversized responses, do not claim a hard byte guarantee. Disclose omissions, truncation and the continuation offset; keep full exports outside model context. Tool discovery and context caching depend on the client; no token-savings guarantee is made.

Legality and data handling

Only submit public pages you are permitted to access and process. Respect site terms, intellectual-property rights and applicable law. Discovering an asset URL does not grant permission to reuse that asset. This independent utility is not affiliated with target websites.

No AI model is used during Actor execution. No paid external API or proxy is required. HTML is processed transiently; it is not saved to the dataset. Dataset rows include supplied/final URLs, page titles and public icon declarations, which may contain query values: do not supply secrets or private URLs. Apify stores run input, datasets and logs under its account retention controls; delete them through Apify when no longer needed. The Actor keeps no cross-run cache.

Failed operations send sanitized diagnostic input, exceptions and actor/build/run IDs to our private GlitchTip service for repair; secret fields and URL queries are removed and reports are retained for 30 days. Apify and the target websites receive ordinary requests. No icon binary is downloaded.

FAQ and troubleshooting

Why is verified always false?

This product discovers declarations, not image availability or correctness. Validate and approve an asset downstream before using it.

Why did a batch fail when rows exist?

At least one selected website failed. Use rows only as partial output, inspect logs and retry the failed source separately after correcting the input or access issue.

Why are there no icons?

The page may have no supported HTML declaration, require JavaScript, or declare icons only in a manifest. With fallback disabled, an empty icon array truthfully records this outcome. A successful website inspection is billed even when it finds no supported declarations. Enable fallback only if an unverified inferred URL is useful.

Can it download, convert or monitor icons?

No. Results contain URL metadata. Schedule and compare datasets externally if recurring enrichment is needed.

Related workflows and support

This is a standalone metadata utility; no additional Actor is required. Use Apify dataset exports or your existing ETL tools for catalogue enrichment. Report bugs and how-to questions through the Actor's Apify Issues tab with public sample input and a run link; never post credentials.