# ScanForAI

> ScanForAI is a free, explainable AI-writing detector: paste any text and see how much of it reads as AI-generated, with the specific signals behind the score. It returns a score plus a per-signal breakdown in plain English — a signal, not a verdict — and is deliberately designed to avoid false accusations.

ScanForAI scores text with stylometric heuristics (sentence rhythm, vocabulary, phrasing habits, predictability, structural scaffolding) combined by a trained classifier. Anonymous scans are never stored. It reads English (including English written by second-language writers), Spanish and Turkish (Turkish text is always read with the Deep scan, 100 words or more); text in other languages comes back "not scored" rather than with a guessed number. The public scanner is at scanforai.com; the logged-in dashboard (scan history, comparisons, API keys) is at app.scanforai.com.

## Pages
- [Scanner (home)](https://scanforai.com/): the free AI-text checker — score, band, and the AI-leaning lines marked in your text; how it works, who it's for, FAQ.
- [Docs](https://scanforai.com/docs/): methodology, interpreting scores, the honest limits, and the API.
- [Accuracy & honesty](https://scanforai.com/accuracy): the bands, the documented failure modes, why a score is a signal and not a verdict.
- [ScanForAI en español](https://scanforai.com/es/): the Spanish scanner and guides — paste Spanish text for a deep scan; the guides cover what to do if you are accused of using AI, how teachers can talk to a student about it, and how far to trust detectors.
- [API reference](https://scanforai.com/docs/api): detect AI text (POST /api/scan) and clean it up (POST /api/cleanup) over HTTP; OpenAPI spec at https://api.scanforai.com/api/openapi.json (Postman-importable), Swagger at https://api.scanforai.com/api/docs.
- [MCP server for AI agents](https://scanforai.com/mcp): Claude Code / Cursor / any MCP client that can send a request header gets five native tools: scan_text, scan_url, cleanup_text, find_hidden_instructions and check_ai_image — `claude mcp add --transport http --scope user scanforai https://api.scanforai.com/api/mcp --header "Authorization: Bearer at_sk_..."` (free key at app.scanforai.com).
- [AI vocabulary](https://scanforai.com/ai-vocabulary): the live, dated list of words and phrases AI models overuse — the same lexicon our detection engine runs on, updated as models change.
- [Detect ChatGPT](https://scanforai.com/detect-chatgpt): what ChatGPT and GPT-5 writing sounds like, watermark/Turnitin/paraphrasing questions answered honestly.
- [Detect Claude](https://scanforai.com/detect-claude): why Claude is the hardest mainstream voice to catch, and what still gives it away.
- [Detect Gemini](https://scanforai.com/detect-gemini): the mixed-document case — Gemini "Help me write" paragraphs inside human drafts, scored per part.
- [Cleanup](https://scanforai.com/cleanup): strip machine typography from AI-written text without rewriting a word — smart quotes, em dashes, and the rare glyphs models love (×, →, ½, ≈); a live tool runs ON the page, free and ungated, no account, always. Uploaded .docx/.odt files come back cleaned with formatting intact (free account).
- [Injection scan — image prompt injection scanner](https://scanforai.com/image-prompt-injection): Reads an image the way a machine would — recovering the text a model would receive from the file, including layers no viewer shows you (very low-contrast text, metadata fields, invisible characters are the well-known examples) — and scores "instruction density" 0-100 across the same low/moderate/elevated/high bands, with a per-read breakdown naming which layer fired. The full set of layers checked is not published. Free with a free account, on the site and over the API (POST /api/injection/image, `at_sk_` key required) — the account is the abuse guard on an upload endpoint, not a paywall. Explicitly NOT an AI-generated-image detector: it asks what an image says TO a model, not whether it was made BY one.
- [Contact](https://scanforai.com/contact): reach the team — bug reports, billing, false-positive reports, school/team rollouts.
- [Browser extension (live on Chrome, Edge and Firefox)](https://scanforai.com/extension): scan any selection, page, link or file from the browser — thin client, engine server-side, nothing stored. Also runs on Brave, Opera, Vivaldi, Arc and other Chromium builds via the Chrome listing; no Safari, and no mobile (browsers don't allow extensions there).
- [For teachers](https://scanforai.com/for-teachers): using detection honestly in the classroom — evidence and conversations, not accusations; non-native cautions.
- [For students](https://scanforai.com/for-students): check your essay before submission; what to do if you're falsely accused of using AI.
- [For writers](https://scanforai.com/for-writers): pre-check drafts, share results without exposing text, the "Checked with ScanForAI" badge. Share links stay live for 180 days.
- [For editors](https://scanforai.com/for-editors): batch-check freelance deliveries and URLs; send writers marked lines instead of a bare score.
- [vs GPTZero](https://scanforai.com/compare-gptzero): honest comparison — the education/second-opinion case, what to do if GPTZero flagged you.
- [vs Originality.ai](https://scanforai.com/compare-originality): honest comparison — credits vs free, the editor workflow.
- [vs CheckForAI](https://scanforai.com/compare-checkforai): CheckForAI has been discontinued (first-party sunset notice) — what it was, and the free alternative.
- [AI detector character limits](https://scanforai.com/ai-detector-character-limits): paste limits compared across detectors — 20,000 characters free here, 50,000 with a free account; verified numbers with dates.
- [Pricing](https://scanforai.com/pricing): the free tier (permanent), Pro at $19/month or $190/year (2 months free — $15.83/month equivalent), Team at $12 per person/month or $120 per person/year (minimum 5 people, set up together with us, higher deep-scan limits than Pro), Enterprise custom. The enforced limits per tier and the honest feature matrix vs GPTZero / Originality.ai / Copyleaks.
- [Privacy Policy](https://scanforai.com/privacy): anonymous scans are never stored.
- [Terms of Use](https://scanforai.com/terms): the accuracy and acceptable-use terms.
- [Refund Policy](https://scanforai.com/refunds): annual plans refundable in full within 14 days of purchase; monthly plans cancel any time and simply don't renew; billing errors always corrected. To ask for a refund, write to us through the contact page and we arrange it.


## Markdown versions (clean text for citation)
Every page below is also served as Markdown at the same path with .md — same content, no navigation or markup.
- https://scanforai.com/accuracy.md
- https://scanforai.com/affiliates.md
- https://scanforai.com/ai-detector-character-limits.md
- https://scanforai.com/ai-vocabulary.md
- https://scanforai.com/cleanup.md
- https://scanforai.com/compare-checkforai.md
- https://scanforai.com/compare-gptzero.md
- https://scanforai.com/compare-originality.md
- https://scanforai.com/contact.md
- https://scanforai.com/detect-chatgpt.md
- https://scanforai.com/detect-claude.md
- https://scanforai.com/detect-gemini.md
- https://scanforai.com/docs/api.md
- https://scanforai.com/docs/index.md
- https://scanforai.com/docs/interpreting-scores.md
- https://scanforai.com/docs/methodology.md
- https://scanforai.com/embed.md
- https://scanforai.com/es/analisis-rapido-o-profundo.md
- https://scanforai.com/es/como-funciona.md
- https://scanforai.com/es/como-hablar-con-un-estudiante-sobre-el-uso-de-ia.md
- https://scanforai.com/es/como-leer-el-resultado.md
- https://scanforai.com/es/como-saber-si-un-texto-esta-escrito-con-ia.md
- https://scanforai.com/es/contacto.md
- https://scanforai.com/es/detectar-chatgpt.md
- https://scanforai.com/es/detectar-gemini.md
- https://scanforai.com/es/detector-de-ia-para-textos-en-ingles.md
- https://scanforai.com/es/embed.md
- https://scanforai.com/es/extension.md
- https://scanforai.com/es/guias.md
- https://scanforai.com/es/index.md
- https://scanforai.com/es/instrucciones-ocultas-en-imagenes.md
- https://scanforai.com/es/para-docentes.md
- https://scanforai.com/es/para-editores.md
- https://scanforai.com/es/para-escritores-y-editores.md
- https://scanforai.com/es/para-estudiantes.md
- https://scanforai.com/es/precios.md
- https://scanforai.com/es/privacidad.md
- https://scanforai.com/es/que-hacer-si-te-acusan-de-usar-ia.md
- https://scanforai.com/es/reembolsos.md
- https://scanforai.com/es/segunda-lengua-y-variedades.md
- https://scanforai.com/es/share-image.md
- https://scanforai.com/es/share.md
- https://scanforai.com/es/son-fiables-los-detectores-de-ia.md
- https://scanforai.com/es/terminos.md
- https://scanforai.com/es/tesis-y-trabajos-academicos.md
- https://scanforai.com/esl-writers.md
- https://scanforai.com/extension-privacy.md
- https://scanforai.com/extension.md
- https://scanforai.com/fast-vs-deep.md
- https://scanforai.com/for-editors.md
- https://scanforai.com/for-students.md
- https://scanforai.com/for-teachers.md
- https://scanforai.com/for-writers.md
- https://scanforai.com/image-prompt-injection.md
- https://scanforai.com/index.md
- https://scanforai.com/mcp.md
- https://scanforai.com/pricing.md
- https://scanforai.com/privacy.md
- https://scanforai.com/refunds.md
- https://scanforai.com/share-image.md
- https://scanforai.com/share.md
- https://scanforai.com/terms.md
- https://scanforai.com/tr/affiliates.md
- https://scanforai.com/tr/ai-detector-character-limits.md
- https://scanforai.com/tr/ai-vocabulary.md
- https://scanforai.com/tr/bir-metnin-yapay-zeka-ile-yazildigini-anlamak.md
- https://scanforai.com/tr/cleanup.md
- https://scanforai.com/tr/compare-checkforai.md
- https://scanforai.com/tr/compare-gptzero.md
- https://scanforai.com/tr/compare-originality.md
- https://scanforai.com/tr/detect-chatgpt.md
- https://scanforai.com/tr/detect-claude.md
- https://scanforai.com/tr/detect-gemini.md
- https://scanforai.com/tr/docs/index.md
- https://scanforai.com/tr/editorler-icin.md
- https://scanforai.com/tr/embed.md
- https://scanforai.com/tr/extension-privacy.md
- https://scanforai.com/tr/extension.md
- https://scanforai.com/tr/fiyatlar.md
- https://scanforai.com/tr/gizlilik.md
- https://scanforai.com/tr/guvenilirlik.md
- https://scanforai.com/tr/hizli-mi-ayrintili-mi.md
- https://scanforai.com/tr/iadeler.md
- https://scanforai.com/tr/ikinci-dilde-yazanlar.md
- https://scanforai.com/tr/iletisim.md
- https://scanforai.com/tr/image-prompt-injection.md
- https://scanforai.com/tr/index.md
- https://scanforai.com/tr/ingilizce-metinler-icin-yapay-zeka-dedektoru.md
- https://scanforai.com/tr/kullanim-kosullari.md
- https://scanforai.com/tr/nasil-calisir.md
- https://scanforai.com/tr/ogrenciler-icin.md
- https://scanforai.com/tr/ogrenciyle-yapay-zeka-kullanimini-konusmak.md
- https://scanforai.com/tr/ogretmenler-icin.md
- https://scanforai.com/tr/rehber.md
- https://scanforai.com/tr/share-image.md
- https://scanforai.com/tr/share.md
- https://scanforai.com/tr/sonucu-nasil-okumali.md
- https://scanforai.com/tr/tezler-ve-akademik-calismalar.md
- https://scanforai.com/tr/yapay-zeka-kullanmakla-suclaniyorsaniz.md
- https://scanforai.com/tr/yazarlar-icin.md
## Key facts
- Fast scan vs Deep scan: fast scans are instant pattern checks (unlimited on every plan); Deep scan trades a few seconds of speed for the highest-quality read of the whole text. Current deep limits: 60/hour anonymous, 30/hour and 300/month with a free account. Results say which method ran.
- Pricing: Free forever · Pro $19/month or $190/year (2 months free) · Team $12 per person/month or $120 per person/year, minimum 5 people, set up together with us, with more deep scans per person than Pro · Enterprise custom (SSO, SLA). Payments are handled by the payment partner named at checkout; charges appear as ScanForAI.
- ScanForAI is operated by Zoniax OÜ, a private limited company registered in Estonia (registry code 16853667).
- The free tier is permanent: the scanner, accounts and history cost nothing. Paid plans add higher limits and are billed monthly or yearly; cancel at any time.
- Not a verdict: every result is a probability estimate from writing patterns and can be wrong in both directions — formal, technical, or non-native human writing can score high, and edited or newer AI can score low.
- Never use a result as the sole basis for a decision about a person, and never to accuse someone.
- Privacy: anonymous scans are never stored or logged; signed-in scans are saved only for your own history and can be deleted.
- Free to use; accounts add scan history and higher limits.