Frequently asked questions
Do you store or see anything I type into the builder?
No. Every generator runs as JavaScript in your own browser. Nothing is sent to a server, there's no account, and no analytics record what you type. Your answers are saved only to your browser's local storage, so they're still there if you close the tab and come back.
Where do I upload the finished files?
Almost all of them belong in the root directory of your website, alongside your homepage — for example example.com/robots.txt. security.txt is the one exception: publish it at /.well-known/security.txt, and optionally also at /security.txt for convenience.
Do I need all 16 files?
No. robots.txt and manifest.json cover the vast majority of practical need. Treat the rest as optional extras: llms.txt, security.txt and ads.txt are worth adding if they apply to you, and the more speculative files are low-cost, low-certainty additions for sites that want to be thorough.
What's the difference between an established, emerging and speculative file?
established means a formal standard or a long-settled convention with real enforcement, like robots.txt or ads.txt. emerging means an active proposal gaining ground but not yet universally supported, like llms.txt or ai.txt. speculative means no governing spec exists at all — a convention a few tools have started using, with no crawler documented to require it.
Will adding these files improve my Google ranking?
Not directly. robots.txt affects whether pages are crawled at all, which matters for indexing, but the AI-facing files here are aimed at large language models and AI answer engines, not classic search ranking. Treat them as part of a wider AI-visibility strategy, not an SEO ranking shortcut.
Does llms.txt guarantee ChatGPT or Claude will cite my site?
No. It's a genuine, actively maintained specification, but independent audits have found no measurable link between having one and being cited more often by AI answer engines, and some major search providers have said they don't use it. Publish it as good practice and a clear content map, not as a guaranteed result.
I already have a robots.txt. Will the builder overwrite my existing rules?
The builder only ever generates a new file for you to download — it can't see or edit anything already on your server. Compare the generated file against your existing one before you upload it, so you don't lose rules you already rely on.
What's included now?
All 16 AI Site Files, 7 schema.org Structured Data types, Image Tools (EXIF/XMP/IPTC viewers, GPS remover, metadata remover, C2PA/AI-metadata remover, OCR), Music Tools (MP3/ID3 viewer, editor, remover), File Tools (Word/Excel metadata viewers, Zip tools, CV Reader, Filetype Determiner, DLL/PE tools, Registry Hive Viewer, and Everyday Formats: vCard, iCalendar, Plist, JWT, JSON and .eml viewers), and a standalone hidden-file finder for images.
Can the JWT Decoder verify any token's signature?
Only HMAC-signed tokens (HS256, HS384, HS512) — paste the secret and it checks the signature locally in your browser. Tokens signed with RSA or elliptic-curve algorithms (RS256, ES256, etc.) are decoded so you can read the header and payload, but their signatures aren't verified; the tool says so plainly rather than showing a false pass or fail.
Is it safe to paste a real JWT or secret into the JWT Decoder?
Nothing you paste is sent anywhere — decoding and signature verification both happen locally in your browser using the Web Crypto API, the same "nothing leaves your browser" promise that applies to every tool on this site. That said, treat any token or secret with the same care you would anywhere else.
Does the .eml viewer let email HTML run scripts or track me?
No. The HTML body of an email is rendered inside a sandboxed frame with scripts disabled and no access to the rest of the page, specifically to prevent anything embedded in the email — tracking pixels included — from running or phoning home automatically.
How are iCalendar recurrence rules decoded into plain English?
The iCalendar Viewer reads the RRULE's frequency, interval, days, month, count and until fields and builds a sentence from them, e.g. "Every 2 weeks on Tuesday and Thursday, until 2025-12-31." Common patterns are covered well; unusually complex or custom rules fall back to showing the raw rule text instead of guessing.
What does the Structured Data section actually give me?
A ready-to-paste JSON-LD snippet for whichever schema.org type applies to a page — Organization, LocalBusiness, Article, Product, FAQPage, BreadcrumbList or Person. Copy it into the <head> of the relevant page; it's what search engines read to build rich results.
Will the GPS remover or metadata remover reduce my photo's quality?
No. Both tools only touch the metadata segments of the JPEG file — the actual pixel data is never re-encoded or recompressed, so image quality is unaffected.
What's the difference between the metadata remover and the C2PA remover?
The metadata remover strips everything — EXIF, XMP and IPTC — indiscriminately. The C2PA remover is targeted: it only removes C2PA Content Credentials manifests and known AI-generation XMP signals, leaving other metadata like your camera details or copyright line untouched.
Can the MP3 tools handle album artwork?
Not yet. The viewer, editor and remover all work with the common text frames (title, artist, album, year, genre, comment) and the legacy ID3v1 tag. Embedded album art is a separate frame type that isn't currently read, written, or preserved when you save changes.
How reliable is the hidden-data finder?
It's genuinely best-effort. It checks for two common, simple techniques — data appended after an image's end marker, and a straightforward sequential least-significant-bit encoding. Encrypted or more sophisticated steganography won't be caught, and a "not found" result never proves a file is clean.
Do the Word and Excel viewers work with the older .doc/.xls formats?
No — only the modern .docx and .xlsx formats, which are zip archives with an internal XML structure the tool reads directly. The older binary .doc/.xls formats use a completely different structure and aren't supported.
Why do some Image Tools show a "speculative" badge?
The same honesty system used for the AI Site Files applies across the whole toolkit. XMP AI-generation signals and the hidden-data finder rely on conventions or heuristics with no governing standard — useful, but not guaranteed. C2PA is further along (an active industry spec) but still marked "emerging" since support isn't yet universal.