usememos/memos · info

not a HTML page

Error message

not a HTML page

What it means

The HTML metadata getter fetched the URL successfully but the response Content-Type was not exactly text/html, so there is no HTML to extract a title/description from. The check is on the parsed mediatype, so parameters like charset are stripped first.

Source

Thrown at internal/httpgetter/html_meta.go:152

}

func GetHTMLMeta(urlStr string) (*HTMLMeta, error) {
	if err := validateURL(urlStr); err != nil {
		return nil, err
	}

	response, err := httpClient.Get(urlStr)
	if err != nil {
		return nil, err
	}
	defer response.Body.Close()

	mediatype, err := getMediatype(response)
	if err != nil {
		return nil, err
	}
	if mediatype != "text/html" {
		return nil, errors.New("not a HTML page")
	}

	htmlMeta := extractHTMLMeta(io.LimitReader(response.Body, maxHTMLMetaBytes))
	enrichSiteMeta(response.Request.URL, htmlMeta)
	return htmlMeta, nil
}

func extractHTMLMeta(resp io.Reader) *HTMLMeta {
	tokenizer := html.NewTokenizer(resp)
	htmlMeta := new(HTMLMeta)

	for {
		tokenType := tokenizer.Next()
		if tokenType == html.ErrorToken {
			break
		} else if tokenType == html.StartTagToken || tokenType == html.SelfClosingTagToken {
			token := tokenizer.Token()
			if token.DataAtom == atom.Body {

View on GitHub (pinned to 14d757ce1f)

Solutions

1. Link to the HTML page that embeds/links the resource, not the resource itself
2. If you own the server, ensure HTML pages are served with `Content-Type: text/html; charset=utf-8`
3. Handle this error gracefully in UI: skip preview generation instead of showing a failure

Example fix

// before
GetHTMLMeta("https://example.com/report.pdf")
// after
GetHTMLMeta("https://example.com/reports/report") // HTML landing page

// Optional pre-check via HEAD (best effort; server may lie)
func looksLikeHTML(u string) bool {
  resp, err := http.Head(u)
  if err != nil { return true } // let the real fetcher decide
  ct := resp.Header.Get("Content-Type")
  return strings.HasPrefix(ct, "text/html")
}

// Not-a-page is expected for non-HTML links: skip the preview silently
if _, err := getter.GetHTMLMeta(u); err != nil {
  if strings.Contains(err.Error(), "not a HTML page") { return nil }
  return err
}

When it happens

Trigger: Fetching a PDF, image, JSON API endpoint, plain-text file, or a download link (Content-Disposition attachments often serve application/octet-stream) through the link-preview path.

Common situations: Pasting a direct link to a PDF, raw GitHub file, image, or API endpoint into a memo; servers misconfigured to serve HTML pages as text/plain; endpoints that content-negotiate to JSON for non-browser clients.

Related errors

• wrong image mediatype
• too many redirects
• internal IP addresses are not allowed
• hostname resolved to no addresses
• invalid URL format

AI-assisted analysis of usememos/memos@14d757ce1f (2026-08-15). Data as JSON: /api/errors/1c3f20edc9e9d5f7. Report an issue: GitHub.