Free website logo extractor
Enter a website and get every logo its page names: the logo in its structured data, its logo pictures, its icons and its share image. The tool runs in the SpecStream workspace after sign-up, on the $10 of credit a new account gets, no card.
What it finds
The tool reads the page’s own HTML and lists each logo with where the page names it, best first. You open the picture or copy its address.
- Structured data logo
- The logo the site states for its organization in schema.org JSON-LD. It is the one the site itself calls its logo.
- Logo pictures
- Each picture on the page whose address, alt text, class or id says logo, and the og:logo tag when the page has one.
- Icons
- The Apple touch icon and every favicon the page links, SVG first, then the largest size.
- Share image
- The og:image the page gives to link previews. It comes last, because it is often a banner and not the logo.
Logo extractor questions
Something we didn't answer? Talk to a human
What does it run on?
One call of the Extract API,
POST /v1/extract, with "include": ["html"]. The call returns the page’s HTML as SpecStream read it, and the tool reads the logos from that HTML.What does a run cost?
A run is one page of the Extract API, at $0.80 per 1,000 pages. It is paid from the $10 of credit a new account gets. A page with no text is not charged.
Can I run it from my code?
Yes. After each run the tool shows its call as curl. With an API key, the same call runs from your own program, for up to 20 websites in one call.
What if a site blocks the read?
The tool says so in one sentence, and the page is not charged.
Read any page from your code
The Extract API returns a page as Markdown or text, with its HTML and links when you ask for them.