Sapling AI detector
Vendor Sapling Intelligence, Inc. Official site sapling.ai/ai-content-detector How this card is sourced
Updated · AI Detector Index editors
Sapling AI Detector checks the first 2,000 characters of a text for free, according to the Sapling AI Content Detector page read 8 November 2026. The vendor is Sapling Intelligence, Inc., a California company, per the Sapling terms page read the same day. Sapling is one of every AI detector we track, on the same fields as the other twenty-one. How this card is sourced.
What Sapling detects
| Field | Value | Source |
|---|---|---|
| Content types detected | Text | sapling.ai/ai-content-detector |
| Generators covered | GPT-5, ChatGPT, OpenAI GPT family, Claude 4.5, Claude 5.5, Gemini 2.5, Qwen3, DeepSeek-V3, Kimi, Llama, Mistral | sapling.ai/ai-content-detector |
| Languages | English, Spanish, Portuguese, French, German, Italian, Dutch, Japanese, Polish, Chinese, Arabic, Russian, Swedish, Korean, Turkish, Romanian | sapling.ai/ai-content-detector |
| Input limits | Web: free check covers the first 2,000 characters; Pro and Enterprise support up to 100,000 characters per check. More accurate after about 50 words. Files: PDF or Word document with selectable text; 5 files per batch (Free), 10 (Pro, Enterprise). API: up to 200,000 characters per request, at least 300 characters recommended; results on text under 500 characters flagged low_confidence; file endpoint accepts .pdf, .docx, .pptx, .odt, .txt, .md, .html or .htm up to 30MB. | sapling.ai/ai-content-detector |
| Free tier | Free plan $0 / month: AI detector checks up to 2,000 characters per check, 5 files per detector batch; count of checks per period not capped. Free 1-month trial of Sapling Pro on registration (no credit card). API: new keys get 50,000 characters/day and 250,000 characters/month. | sapling.ai/pricing |
| API | Yes, sapling.ai/docs/api/detector | sapling.ai/docs/api/detector |
| Integrations | Chrome extension AI Content Detector for ChatGPT, Word add-in, Browser extensions (Chrome, Firefox, Edge, Opera, Brave) for the writing product, Google Docs, Gmail, Outlook, 73 integrations including Zendesk, Salesforce, Slack, Notion, WordPress, MCP server | sapling.ai/integrations |
| Data retention | Privacy Policy: When you use our Services, we collect text entered or otherwise transmitted by you in connection with your use of the Services. Your content is stored by us and used only to improve the Service. We do not sell or rent customer data. API: AI detection results are cached on the entire text for 3 days; send zero_retention: true to score a text without Sapling keeping a copy of it; teams with data retention disabled get this behavior on every request. Servers on AWS us-west. | sapling.ai/privacy |
| Ownership | Sapling Intelligence, Inc. (California); developed by former researchers at Berkeley, Stanford, Meta, Google. Parent company not published. | sapling.ai/tos |
| Independent studies | 3 records | Studies index |
Sapling AI Detector scans text only. The vendor names GPT-5, ChatGPT and the OpenAI GPT family, Claude 4.5 and 5.5, Gemini 2.5, Qwen3, DeepSeek-V3, Kimi, Llama and Mistral as covered generators, per the detector page read 8 November 2026. Sixteen languages are listed on the same page: English, Spanish, Portuguese, French, German, Italian, Dutch, Japanese, Polish, Chinese, Arabic, Russian, Swedish, Korean, Turkish and Romanian.
On humanized text, Sapling's changelog entry for 17 September 2026 reports "significantly higher detection of humanized AI text", with detection of humanized AI text up by 3.9% absolute (Sapling detector API docs, read 8 November 2026). The same docs say that small modifications to AI-generated text can cause that text to no longer be flagged.
Sapling's privacy policy says: "Your content is stored by us and used only to improve the Service" (Sapling privacy policy, read 8 November 2026). The API docs add a zero_retention option that scores a text without Sapling keeping a copy, and results are otherwise cached for 3 days; the card leads with the privacy policy because retention is the default and zero retention is opt-in. Integrations include a Chrome extension, a Word add-in, Google Docs, Gmail, Outlook and an MCP server (Sapling integrations page, read 8 November 2026). Two related cards: GPTZero and Originality.ai, both scored alongside Sapling in Walters 2023.
Limits and plans
| Plan | Price | Limits | Checked on |
|---|---|---|---|
| Pro (monthly) | $25 / month | Advanced writing assistance and AI detector checks up to 100,000 characters per check, 10 files per batch, email and ticket support. | |
| Pro (annual) | $12 / month, billed annually | Same as Pro (monthly). | |
| Enterprise | Custom ... Starts at 10 seats, $15/seat/month | 100,000 characters per check, team controls, SSO/SCIM, self-hosted, dedicated support. | |
| API | From $0.005 per 1,000 characters, with volume discounts; One-time $5 minimum | Calculator: 10 million chars x $0.005/(1k chars) + 20 million chars x $0.00375/(1k chars); over 100 million chars contact us; subscribed keys default 10 million characters per day and 25 million per month. |
Web checks cover the first 2,000 characters on the Free plan and up to 100,000 characters per check on Pro and Enterprise, with 5 files per detector batch on Free and 10 on Pro, per the detector page read 8 November 2026. Sapling says results are more accurate after about 50 words. The API accepts up to 200,000 characters per request, recommends at least 300, and marks results on text under 500 characters as low_confidence; the file endpoint takes .pdf, .docx, .pptx, .odt, .txt, .md, .html or .htm up to 30MB (detector page and API docs, read 8 November 2026).
Prices are from the Sapling pricing page read 8 November 2026. The free tier is a $0 per month plan with each check capped at 2,000 characters and no cap on the number of checks. Registration brings a free 1-month trial of Sapling Pro with no credit card, and new API keys get 50,000 characters per day and 250,000 per month (pricing page, read 8 November 2026).
What the vendor claims
Accuracy varies with the dataset, text length, and decision threshold. Measure false positives and missed detections on examples from your workflow. (No single accuracy number published.)
Sapling says: "Accuracy varies with the dataset, text length, and decision threshold. Measure false positives and missed detections on examples from your workflow." On false positives, Sapling says that on internal held-out evaluations human-text false positives fell by 60% with the 17 September 2026 model upgrade on the modern-text set. Both statements are on the Sapling AI Content Detector page (sapling.ai/ai-content-detector, read 8 November 2026); the 60% is a relative change against Sapling's previous model.
Sapling also says: "No current AI content detector (including Sapling's) should be used as a standalone check to determine whether text is AI-generated or written by a human" (same page, read 8 November 2026).
What independent studies found
- AIDI-S-0002Liang 20232023
Sapling was one of seven detectors tested (accessed March 15, 2023). Across the seven, the paper reports an average false positive rate of 61.22% on 91 human-written TOEFL essays and near-perfect accuracy on 88 US 8th-grade essays. Per-detector rates appear only in a bar chart (Fig. 1a) and are not stated as numbers.
- AIDI-S-0003Walters 20232023
Sapling: 65% of 126 documents correct, 7% incorrect, 28% uncertain; 93% correct on 42 GPT-3.5 papers, 33% on 42 GPT-4 papers; 69% of 42 human papers correct, 0 incorrect, 31% uncertain. Tested June 25 to July 12, 2023.
- AIDI-S-0023Kar 20242024
Read one 500-word ChatGPT 3.5 article as 100% AI and held 100 AI on all three paraphrases (QuillBot, Grammarly, ChatGPT) (Table 1, single text).
Liang 2023 included Sapling among seven detectors accessed 15 March 2023. Across the seven, the paper reports an average false positive rate of 61.22% on 91 human-written TOEFL essays and near-perfect accuracy on 88 US 8th-grade essays; per-detector rates appear only in a bar chart (Liang 2023 PDF, read 8 November 2026).
Walters 2023 ran Sapling against 126 documents between June 25 and July 12, 2023: 65% correct, 7% incorrect and 28% uncertain. Sapling caught 93% of 42 GPT-3.5 papers and 33% of 42 GPT-4 papers. On 42 human papers, 69% were correct, 0 were labelled AI and 31% came back uncertain (Walters 2023 PDF, read 8 November 2026).
Sapling's zero false positives in Walters 2023 arrived with 31% of human papers returned as uncertain, so the clean score was bought partly by declining to answer. A free check also reads only the first 2,000 characters, so on a long document the score says nothing about the rest.
Kar 2024 joined the list on 9 November 2026 and took the count to 3: Sapling read one 500-word ChatGPT 3.5 article at 100% AI and held 100 AI through QuillBot, Grammarly and ChatGPT paraphrases, a single-text observation that moves no rank field.
Where Sapling appears in the index
Rankings
- AIDI-R-01AI detectors ranked by independent testing Two records added in the fifth wave on 9 November 2026, Pegoraro 2023 and Fleckenstein 2024, changed nine counts and none: Pegoraro 2023 tested ZeroGPT, GPTZero, Copyleaks, Originality.ai, Writer, Content at Scale (BrandWell), DetectGPT, the GPT-2 Output Detector and GLTR, lifting GPTZero to 12, the GPT-2 Output Detector to 10, Copyleaks to 7 and Originality.ai and ZeroGPT to 6; Fleckenstein 2024 tested 289 teachers and no software, so it counts for no card ([Fleckenstein 2024 record](/studies/fleckenstein-2024-teachers/)). Illuminarty, one of four cards added in the same wave, joined with its one image record from Ha and Passananti 2024, so 18 of 72 cards now qualify.
- AIDI-R-02AI detector false positive rates, ranked
- AIDI-R-04AI detectors ranked by languages supported
- AIDI-R-05AI detectors ranked by maximum input per check
- AIDI-R-06AI detectors ranked by stated data retention
- AIDI-R-11AI detectors ranked by named integration surfaces
- AIDI-R-12AI detectors ranked by the upload formats the vendor states
- AIDI-R-14AI detector APIs ranked by published price per 1,000 words
- AIDI-R-17AI detectors ranked by the generator models the vendor names
- AIDI-R-18AI detectors ranked by study records on non-English and non-native text
Fits these use cases, questions and comparisons
- AIDI-U-01AI detectors for business and managers
- AIDI-U-03AI detectors for copywriters and agencies
- AIDI-U-04AI detectors for job applications and recruiters
- AIDI-U-06AI detectors for admissions offices
- AIDI-U-07AI detectors for marketplaces and review platforms
- AIDI-U-08AI detectors for researchers checking their own manuscripts
- AIDI-U-09AI detectors for law firms and courts: an index fit check
- AIDI-U-10AI detectors for grant and proposal reviewers
- AIDI-U-12AI detectors for localisation and translation teams
- AIDI-U-13AI detectors for government and public sector buyers
- AIDI-U-14AI detectors for self-publishing authors: the cards that take a manuscript and say what happens to it
- AIDI-U-15AI detectors for customer support quality assurance: the cards with an API, a short-text statement, a bulk route and a retention policy
- AIDI-C-15Sapling vs GPTZero AI detector: two text detectors with three shared studies, on the same 14 fields
Sapling appears in three rankings. AI detectors ranked by independent testing lists two studies for Sapling, Liang 2023 and Walters 2023. AI detector false positive rates carries 0 of 42 human papers labelled AI from Walters 2023, with the 31% uncertain share noted. AI detectors by languages supported places Sapling at 16 languages, per the detector page read 8 November 2026.
Sapling sits on three audience views: AI detectors for business and AI detectors for copywriters and agencies, because of the metered API, and AI detectors for job applications, because of the free check. The detector hub lists all twenty-two cards.
Sapling leads the free API question, added 9 November 2026, on 50,000 characters a day and 250,000 a month for a new API key, is the one card tabled in characters on the free checks per day ranking, and has one record on the non-English evidence ranking, the Liang 2023 TOEFL essays.
Questions about Sapling AI detector
How much text does Sapling check for free?
The first 2,000 characters of each check. Sapling's Free plan costs $0 per month, caps each check at 2,000 characters, allows 5 files per batch and sets no cap on how many checks you run, per the pricing page read 8 November 2026. Pro raises the cap to 100,000 characters per check.
What did Walters 2023 find for Sapling?
Sapling classified 65% of 126 documents correctly, 7% incorrectly and 28% uncertain in Walters 2023. On the 42 human papers, 0 were labelled AI and 31% came back uncertain; on the 42 GPT-4 papers, 33% were caught. The test ran June 25 to July 12, 2023.
Does Sapling have an API, and what does it cost?
Yes. Sapling says the API starts at $0.005 USD per 1,000 characters with a one-time $5 minimum, accepts up to 200,000 characters per request, and offers a zero_retention option, per the pricing page and API docs read 8 November 2026.
How many languages does Sapling's detector cover?
Sixteen. The Sapling detector page read 8 November 2026 lists English, Spanish, Portuguese, French, German, Italian, Dutch, Japanese, Polish, Chinese, Arabic, Russian, Swedish, Korean, Turkish and Romanian, which places Sapling at 16 on the languages ranking.
Sources and revisions
Sources
- sapling.ai/ai-content-detectorRead 8 November 2026
- sapling.ai/tosRead 8 November 2026
- sapling.ai/pricingRead 8 November 2026
- sapling.ai/docs/api/detectorRead 8 November 2026
- sapling.ai/privacyRead 8 November 2026
- sapling.ai/integrationsRead 8 November 2026
- arxiv.org/pdf/2304.02819v3Read 8 November 2026
- degruyterbrill.com/document/doi/10.1515/o...Read 8 November 2026
Revision log
- Card created2026-11-09: kar 2024 record added in the sixth wave, 3 study records2026-11-09: free api question, free checks per day ranking and non-english evidence ranking linked (seventh wave)
). Illuminarty, one of four cards added in the same wave, joined with its one image record from Ha and Passananti 2024, so 18 of 72 cards now qualify.: register plate](/img/independently-tested-480.webp)
