Two things drive most Walter Writes cancellations, and neither is output quality. The first is the monthly word pool that resets to zero whether you used it or not. The second is a refund window that closes after three days and 1,500 words. Both are documented in their own help centre, and both are worth checking on whatever you move to next.
Paste some text or load the example to estimate AI likelihood.
Their help centre is public, which is more than several competitors manage, so these figures come straight from it. All checked 16 August 2026.
Monthly word pools. Starter is "30,000 words per month total", Pro is "70,000", Elite is "200,000", Team is "500,000" (support.walterwrites.ai, "How do Walter Writes AI subscription plans work?").
No rollover. "No, unused word credits don't carry over. Your word balance resets at the start of each new billing cycle." The same article states this applies to monthly and annual plans alike.
A hard stop, not a slowdown. "If you reach your monthly word limit, access will be paused until your next billing cycle starts." Upgrading mid-cycle is prorated. Team is described as the only plan that can buy extra credits after running out.
Detector scans are free of the meter. "Scanning your content with our AI Detector doesn't use any of your subscription credits. Only the words processed through the Humanizer count toward your monthly or annual word balance." A genuinely good design decision, worth naming.
A 50-word floor. "Our platform requires a minimum of 50 words to process any text through the Humanizer", and short submissions do not deduct from your balance.
Free trial. 300 Humanizer words plus "full access to the AI Detector for 3 days", email sign-up, no credit card. It does not refresh.
Refunds. Eligibility requires all three of: a request within "3 days of signing up", usage of "1,500 words or less", and that it is your "first-ever subscription". The article states the company does not refund for dissatisfaction with results, for failed detection tests, or for missed cancellations, and that "All subscription renewals are final and non-refundable". Requests go to support@walterwrites.ai with a stated response time of 2 business days.
Languages. 30 processing languages are listed: Arabic, Bengali, Chinese, Czech, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Korean, Norwegian, Persian, Polish, Portuguese, Romanian, Russian, Spanish, Swedish, Thai, Turkish, Ukrainian, Vietnamese. The dashboard interface itself is available in eight: English, Spanish, French, German, Portuguese, Italian, Chinese and Dutch.
We could not verify prices. Requests to walterwrites.ai/pricing returned HTTP 403 to every automated check on 16 August 2026, and their help centre says only that "Our pricing page has a toggle so you can see the cost for each option", with yearly paid upfront. Read the live page before committing.
Thirty languages. Our detector is English-only. If you write in Vietnamese, Bengali or Ukrainian, this comparison ends here and Walter Writes wins it.
Detection that does not eat your budget. Their AI Detector runs free of the word meter. If your workflow is scan-heavy and rewrite-light, that is a genuinely favourable structure.
A public help centre with real answers. You can read their refund policy, per-request limits and rollover policy without an account. Two competitors in this category blocked every attempt to read their pricing page at all. Publishing terms openly is a mark in Walter Writes' favour.
Configurable output. Their getting-started guide describes three controls: a Readability level, a Purpose setting for academic or professional contexts, and a Detection Bypass Level governing how aggressively text is rewritten.
And the boundary that matters most. Walter Writes sells a humanizer. Their own guide is blunt about what that means: "Walter rewrites for detection, not for meaning", followed by the instruction to always review the output. TextSight does not sell a humanizer. Our rewriter targets clarity, and we would rather lose the sale than let you buy the wrong category of product.
Their "Word Per Request Limit Exceeded" article lists per-request ceilings of 750 words on Starter, 1,500 on Pro, and 2,000 on Elite and Team. Their "Getting Started with the Walter Writes AI Humanizer" article says "Your word limit depends on your plan: 500 words (Starter), 1,200 (Pro), and 1,700 (Unlimited)", using a plan name, Unlimited, that does not appear in the plans article at all.
Both were live on 16 August 2026. We do not know which set is current. If per-request size matters to you, ask their support to confirm before subscribing.
Whatever you move to, get answers to these six before you enter a card.
1. What is metered: words, requests per day, or documents? Walter Writes meters humanizer words monthly. StealthWriter meters runs per day. We meter checks per day. 2. Does unused allowance roll over? Walter Writes says no. Ask anyway. 3. What is the per-request ceiling, as opposed to the monthly total? Different numbers, sometimes published inconsistently. 4. What happens at the limit: hard pause, overage charge, or throttle? 5. What is the refund window, in days and in usage? Three days and 1,500 words is narrow. Some are narrower. 6. Is the pricing page readable without an account? If a vendor will not publish plan terms openly, you are agreeing to something you have not read.
| Walter Writes | TextSight | |
|---|---|---|
| Metering | Monthly humanizer words: 30k Starter, 70k Pro, 200k Elite, 500k Team | Checks per day: 3 anonymous, 60 free account, 200 Starter, unmetered Pro and above |
| Rollover | None; balance resets each cycle | Daily reset |
| Per-request size | 750 / 1,500 / 2,000 in one article; 500 / 1,200 / 1,700 in another | 1,000 words anonymous, 5,000 words signed in |
| Minimum input | 50 words | 25 words |
| Detection cost | Free, does not consume credits | Included in the check allowance |
| Free entry | 300 humanizer words, detector for 3 days, no card | 3 checks/day forever with no account; 60/day free account |
| Refunds | 3 days, 1,500 words, first subscription only; renewals final | Not something we publish terms for |
| Languages | 30 processing, 8 interface | English-only detector |
| Price | Not verified (pricing page returned HTTP 403) | Starter $9.99 / $7.49 annual, Pro $19.99 / $14.99, Business $39.99 / $29.99, Enterprise $79.99 / $59.99 |
All Walter Writes figures from support.walterwrites.ai, checked 16 August 2026.
Our detector is an English-only ensemble of DeBERTa, RoBERTa and ELECTRA. Under 30 reads as "likely human-written", 30 to 69 as "mixed signals", 70 and above as "likely AI-generated". Sentence-level analysis and PDF export start at Starter. Chrome extension and REST API start at Pro. Teams and branded reports start at Business. File upload, URL scan and bulk scan sit across the paid plans.
Given that Walter Writes serves 30 languages, one caveat is especially relevant to its users: our model carries a known false-positive bias against writers who learned English as a second language. If you drafted in Hindi or Portuguese and translated into English, expect elevated scores that may say nothing about how the text was produced.
A writer working in a language other than English. Walter Writes, without hesitation.
Someone whose only complaint was the word pool running out. Compare per-day metering instead of per-month. StealthWriter's daily model may suit you better than any monthly pool.
Someone who used the free detector far more than the humanizer. You were using the free half of the product. Our free account gives 60 checks a day at 5,000 words with saved history, and Starter at $9.99/mo adds sentence-level analysis.
An editor or agency vetting delivered copy. Bulk scan, PDF export at Starter, branded reports at Business.
A student worried about a false accusation. Read the ESL bias note above, then use sentence-level analysis to see which passages score and why. No detector output is proof.
All checks performed on 16 August 2026.
TextSight pricing, quotas and features are our own published figures.
The deeper head-to-head: feature table, where GPTZero wins, where TextSight wins, and migration steps.
Read the comparePerplexity, burstiness, and transformer scoring explained, so you can read any detector's verdict critically.
Read the guideWhy detectors over-flag ESL and formal writing, and how sentence-level evidence helps you defend real work.
Read the guideThe full ranking with detection approach, pricing, and use-case fit side by side.
See the rankingStart with TextSight's free tier. No card, no signup, no commitment. Your first scan in about six seconds, with the evidence shown line by line.
Honest comparisons vs other tools.