JustDone AI detector
Vendor GM Appdev Limited Official site justdone.com/ai-detector How this card is sourced
Updated · AI Detector Index editors
AI Detector by JustDone checks up to 15,000 words in one scan, according to the JustDone AI Detector page read 8 November 2026. The vendor is GM Appdev Limited, Office 202, 2nd Floor, 15 Nafpliou Str., 3025, Limassol, Cyprus, as stated on the same page. JustDone is one of twenty-two cards in the detector register, each built on the same fields. How this card is sourced.
What JustDone detects
| Field | Value | Source |
|---|---|---|
| Content types detected | Text | justdone.com/ai-detector |
| Generators covered | ChatGPT, GPT-3.5, GPT-4, GPT-5, Claude, Gemini, Llama, DeepSeek, additional open-source and commercial LLMs | justdone.com/ai-detector |
| Input limits | Paste text directly or upload a PDF, Word, or TXT file. Check up to 15,000 words in one scan (Premium: unlimited AI checks up to 15,000 words). Minimum text length for reliable results: 350 characters. File size not published. | justdone.com/ai-detector |
| Free tier | No setup and no account needed for your first scan; quantity per period not published | justdone.com/ai-detector |
| Integrations | Chrome Extension | justdone.com/ai-detector |
| Data retention | Privacy Policy (Effective 25.06.2025), section 11: We will store your personal data for as long as it is reasonably necessary for achieving the purposes set forth in this Privacy Policy (including providing the Service to you). We also may collect personal information that is included in the User Content. Detector page: Your text is never sold or published. Whether submitted text trains models: not published. | justdone.com/legal/privacy-policy |
| Ownership | GM Appdev Limited, Office 202, 2nd Floor, 15 Nafpliou Str., 3025, Limassol, Cyprus. Further parent not published. | justdone.com/ai-detector |
AI Detector by JustDone scans text only. The detector page FAQ mentions analysing "images, text, keywords", but no image detection product exists on the site, so the card records text as the single modality (detector page, read 8 November 2026). JustDone names ChatGPT, GPT-3.5, GPT-4, GPT-5, Claude, Gemini, Llama, DeepSeek and additional open-source and commercial LLMs as covered generators.
On edited and humanized text, JustDone says the detector "keeps learning from new AI writing" that has "already been run through a paraphraser or grammar tool", and lists "edited-AI traces" as a feature so that "humanized" AI text is still caught. JustDone's blog states the detector correctly flags 60% of AI-edited human drafts (detector page, read 8 November 2026). JustDone also shows a Turnitin estimate and says: "Estimation based on our own model analysis. JustDone is not affiliated with Turnitin, and institutional results may differ."
JustDone's privacy policy, effective 25 June 2025, says in section 11: "We will store your personal data for as long as it is reasonably necessary for achieving the purposes set forth in this Privacy Policy (including providing the Service to you)" (JustDone privacy policy, read 8 November 2026). The detector page adds that submitted text is never sold or published. The one listed integration is a Chrome Extension. Two related cards: Turnitin, whose score JustDone estimates, and Scribbr, the other free detector in the index aimed at student writing.
Limits and plans
JustDone accepts pasted text or an uploaded PDF, Word or TXT file, scans up to 15,000 words at once, and asks for at least 350 characters for reliable results (detector page, read 8 November 2026). Premium is described as unlimited AI checks up to 15,000 words per scan.
JustDone publishes no prices the index could confirm: the JustDone pricing page renders its figures client-side, and the card carries none (read 8 November 2026). The free tier needs no setup and no account for the first scan, per the detector page read 8 November 2026; the free quantity per period is not stated on the pages read, so the free tier has no period on this card.
What the vendor claims
80% overall detection accuracy across all text types; 98% accuracy on academic and scientific texts; 10.3% error rate, published openly, because no honest detector claims 100% (labelled Independently benchmarked). Blog (updated March 19, 2026): AI detection accuracy on pure AI text 80%, total error rate 10.3%, full correct classification rate 67.5%.
JustDone says: "80% overall detection accuracy across all text types; 98% accuracy on academic and scientific texts; 10.3% error rate, published openly, because no honest detector claims 100%." The figures are labelled "Independently benchmarked" on the detector page (justdone.com/ai-detector, read 8 November 2026). JustDone's blog, updated 19 March 2026, gives the same evaluation as 80% on pure AI text, a 10.3% total error rate and a 67.5% full correct classification rate.
Two of JustDone's pages describe the academic result differently: the detector page says 98% on academic and scientific texts, and the blog says "almost 100%" with ArXiv 100%, News 100% and PubMed 83.3%. The card uses the detector page figures, 80% overall, 98% academic and 10.3% error, because the detector page is the product claim, and summarises the blog table beside them. The "Independently benchmarked" label is quoted as JustDone's own words; the blog describes JustDone's own evaluation framework and production API testing, and no third-party study of JustDone exists in the index.
JustDone also says: "Every AI detection result is a probability, not a verdict" and "No detector is completely accurate and can make errors in both directions" (detector page, read 8 November 2026).
What independent studies found
No study in the index has tested JustDone; the detectors that have been tested are ranked on AI detectors ranked by independent testing.
JustDone's 10.3% error rate is the vendor's own number from the vendor's own test set, labelled "Independently benchmarked" by the vendor. A label is not a study, and the index treats the figure as a claim until a published paper puts JustDone on a shared sample.
Where JustDone appears in the index
Rankings
- AIDI-R-05AI detectors ranked by maximum input per check
- AIDI-R-11AI detectors ranked by named integration surfaces
- AIDI-R-12AI detectors ranked by the upload formats the vendor states
- AIDI-R-13AI detectors ranked by minimum text length per check
- AIDI-R-17AI detectors ranked by the generator models the vendor names
Fits these use cases, questions and comparisons
- AIDI-U-08AI detectors for researchers checking their own manuscripts
- AIDI-U-10AI detectors for grant and proposal reviewers
- AIDI-U-11AI text detectors for newsrooms and fact desks
- AIDI-Q-01Is there a 100% accurate AI detector? What published tests show
- AIDI-Q-04Is JustDone's AI detector legit or accurate? What is published
The four rankings order cards by study count, published false positive rate, free words per month and language count. JustDone's record carries no value in any of those four fields, so no ranking lists JustDone; the detector hub lists JustDone with the fields it does carry.
JustDone sits on one audience view: AI detectors for job applications, because the first scan is free without an account and the detector is text only.
Questions about JustDone AI detector
How accurate does JustDone say its AI detector is?
80% overall, 98% on academic and scientific texts, with a 10.3% error rate, per the JustDone detector page read 8 November 2026. JustDone's own blog, updated 19 March 2026, reports the same evaluation as 80% on pure AI text and 67.5% full correct classification. The figures come from JustDone's own test set.
Is JustDone's AI detector free?
The first scan is free with no setup and no account, per the detector page read 8 November 2026. JustDone publishes no prices the index could confirm, because the pricing page renders its figures client-side, so the card carries no plan table.
How much text can JustDone check at once?
Up to 15,000 words in one scan, pasted or uploaded as PDF, Word or TXT, per the detector page read 8 November 2026. JustDone asks for at least 350 characters for reliable results and says short fragments, heavily formatted writing and lists are the hardest cases.
Does JustDone's Turnitin estimate predict a Turnitin score?
No. JustDone says the Turnitin estimate is "based on our own model analysis", that JustDone "is not affiliated with Turnitin", and that institutional results may differ, per the detector page read 8 November 2026. The Turnitin card carries Turnitin's own published rates.
Sources and revisions
Sources
- justdone.com/ai-detectorRead 8 November 2026
- justdone.com/legal/privacy-policyRead 8 November 2026
- justdone.com/pricingRead 8 November 2026
