About Us - TextSight.ai | AI Content Detection Made Actionable

Our Mission

Detection that tells you what to do next.

AI-generated content is everywhere. From student essays to marketing copy, distinguishing between human and AI writing has become one of the most important challenges of our time.

TextSight.ai was created to solve this problem — not just by detecting AI content, but by giving users actionable insights to improve their writing. We go beyond simple detection with our Authenticity Score and AI-powered rewrite suggestions.

Our goal is simple: make AI content detection accessible, accurate, and genuinely useful for everyone.

Why TextSight.ai?

Most AI detectors tell you something is AI-generated and stop there. TextSight.ai tells you why and shows you how to fix it. Our Authenticity Score gives you a clear path from detection to improvement — making it a complete workflow that helps you write better.

Our story.

TextSight.ai was founded in 2025 by Dipak Bhosale with a clear vision: to build an AI content detection tool that doesn't just flag problems — it helps solve them.

As AI writing tools like ChatGPT, Claude, and Gemini became mainstream, the need for reliable detection grew rapidly. Educators needed to verify academic integrity. Businesses needed to ensure content authenticity. Writers wanted to understand where their voice sounded too "machine-like."

But existing detectors were limited — they'd give you a percentage and leave you guessing. There was no guidance on what to do next. That's the gap TextSight.ai was built to fill.

"A percentage isn't an answer. Which sentence is the answer. And the next answer is, here's how you'd fix it."

What we believe.

Detection without remediation just creates fights. So we built both in one workspace — the detector flags it, the Authenticity Score quantifies it, and a one-click rewrite closes the loop. No tab-switching, no second tool, no guessing.

Where we are now.

Starting as a solo project, TextSight.ai has grown into a comprehensive platform trusted by users worldwide. From the Authenticity Score to AI-powered rewrite suggestions, every feature was designed with one principle: detection should be actionable, not just informational.

TextSight.ai is operated by Lacewing Technologies, an independent company based in Maharashtra, India. We're not owned by, funded by, or affiliated with any of the AI model providers (OpenAI, Anthropic, Google, Meta) whose output we detect — that independence is the reason we can publish honest accuracy numbers and known limitations openly.

Where we're going.

Towards the version of TextSight where you stop noticing the detector and just write better. Where the score on the side moves the way Grammarly's underlines do — fast, accurate, and quietly. Where 20 tools become 50. Where authenticity preserves not just facts but your voice.

If that sounds like something you want to use — or help build — we'd love to hear from you.

By the numbers

Honest miles. Real claims.

Up to 99.2%*

Detection accuracy
internal benchmark

50+

Languages supported

2025

Founded

Independent

No outside investors

What we believe

Six principles, no exceptions.

Every product decision, pricing change, and policy update gets weighed against these six rules. None of them are negotiable, and none of them shift when a competitor ships a flashier number.

Show your work

Per-sentence scores, model attribution, exportable PDFs. No vibes-check claims, no single-percentage black box. If we say something is AI, we point at it and tell you which signal triggered the flag — burstiness, perplexity, vocabulary fingerprint, or stylometric variance.

Your text is yours

Zero retention by default. Never stored without your opt-in. Never used to train any model — ours or anyone else's. Never seen by humans on our side. Region-pinned storage available for enterprise customers under DPA, and one-click deletion of any scan or your full account at any time.

False positives are the worst thing we can do

We obsess over them. Up to 99.2% accuracy on our internal benchmark, with a false-positive rate kept below 0.8% on 4.2 million human-writing samples. AI detection is probabilistic, so we publish our methodology, our known limitations, and a free human-review appeal path on every paid plan.

Ship the rewrite, not just the score

A flag without a fix is a dead end. Every score sits next to an Authenticity Score and a one-click rewrite path with three tone presets, fact-locking on numbers and citations, and a closed-loop verifier that re-scores until the result reads natural. No other detector ships the fix in the same product.

Honest pricing, honest claims

Four published tiers with every limit and feature listed in one table. No hidden character meters, no surprise overage fees, no sales call to see a number. When we miss an accuracy mark, we publish the regression on the methodology page. Marketing copy never moves a decimal point that the benchmark does not back up.

Accessible to anyone who writes

The free tier is the full product on a daily quota, not a stripped feature set. Students checking a single essay, hobby bloggers, and small-team editors get the same engine, the same Authenticity Score, the same sentence highlights as a Pro user. Upgrade only when daily volume forces it.

What we build

What TextSight builds and why.

Six pillars of the product, all built around one belief: detection is only useful when paired with a path to fix what gets flagged.

🌍

Multi-family AI detection

Calibrated across GPT-4, Claude 3.x, Gemini, Llama, and Mistral outputs. Many tools score only on GPT data and miss content written by Claude or Gemini. TextSight is trained on a broader corpus so the verdict holds regardless of which model wrote the draft.

🎯

Sentence-level reasoning

Not a black-box percentage. Every flagged line is highlighted with the signal that triggered it, so writers can audit, defend, or rewrite with confidence. That visibility is what lets a student contest a false positive or an editor surgically rewrite three sentences instead of redrafting an entire article.

✨

Integrated AI rewriter

A one-click rewrite tuned for ethical AI assistance. Meaning, facts, and citations are preserved while register, sentence rhythm, and vocabulary are pulled back toward natural human writing. Designed for writers polishing AI drafts, not for working around legitimate academic checks.

💰

Transparent pricing

Four published tiers: Free, Starter, Pro, and Business. No hidden character meters, no surprise overage fees, no sales call to see a number. The pricing page lists every limit and every feature in one table so the right plan is obvious before checkout.

🔌

Extension, plugin, and API

Chrome extension for right-click scanning on any page, a WordPress plugin for editors checking posts before publishing, and a documented REST API for teams wiring detection into their own workflows. Same model, same scores, three surface areas.

🏢

Independent, founder-led

TextSight.ai is operated by Lacewing Technologies, founded by Dipak Bhosale in 2025. We are not owned by, funded by, or affiliated with any of the AI model providers whose output we detect. That independence is the reason we can publish honest accuracy numbers and known limitations openly.

Made by

Founder-led. Independent.

Operated by Lacewing Technologies · based in Maharashtra, India · serving customers globally.

Dipak Bhosale

Founder & CEO

Built TextSight from a solo project in 2025. Reads every support email. Reachable at contact.html.

Hiring

Engineers & linguists

We're growing the team. If you're excited by AI-content authenticity work, get in touch.

Company FAQ

Frequently asked about TextSight as a company.

Eight questions we get from prospective customers, journalists, and partners.

Who built TextSight?

TextSight was founded in 2025 by Dipak Bhosale and started as a solo project. The original motivation was personal frustration with existing detectors that returned a verdict without explaining which sentences caused it. The first shipped version focused on closing that loop with sentence-level highlights, an Authenticity Score, and a one-click rewrite path in a single workflow. The team has since grown but the founding principle is the same: detection should hand the writer something they can act on, not just a number.

Where is the company based?

TextSight.ai is operated by Lacewing Technologies, a company based in India. We serve customers globally and run our infrastructure on cloud providers with regional capacity, so users see the same product whether they sign up from North America, Europe, Asia-Pacific, the Middle East, or Latin America.

What makes TextSight different from competitors?

Three things. First, the detector is calibrated across multiple model families instead of being tuned only on GPT outputs, so it does not silently fail on Claude or Gemini content. Second, every score comes with sentence-level highlights that show which lines triggered the verdict, which makes the result auditable rather than a black box. Third, the AI rewriter ships inside the same product, so the path from "this is flagged" to "this reads naturally" is a single click instead of three subscriptions across three vendors.

What are the customer support hours and channels?

Email support runs Monday to Friday with a target first response under 24 hours for all paid plans and best-effort response for free users. In-app chat is available during India business hours and triages to email outside that window. Pro and Business plans get priority routing on the same channels. The contact page lists current addresses and the in-app help center has self-serve articles for the most common questions.

What is the refund policy?

Monthly plans can be cancelled at any time from the billing page and access continues until the end of the current billing period. We offer a refund for the most recent monthly charge if a user contacts support within 7 days of payment and has not consumed a substantial portion of the included quota. Annual plans are eligible for a prorated refund of unused months within 30 days of purchase. The full policy lives on the terms page.

How is data privacy handled and where is data stored?

Submitted text is processed for the scan or rewrite the user requested and is never used to train, fine-tune, or improve any AI model, ours or anyone else's. Free-tier text is discarded after processing; paid-plan history is retained for the documented retention window on each tier so users can review past scans, and any scan or the entire account can be deleted on demand. Data is encrypted in transit and at rest, and infrastructure runs on major cloud providers with industry-standard security controls. The privacy and content-handling pages document the retention table in full.

Is TextSight VC-funded or bootstrapped?

TextSight is independently operated and is not owned by or affiliated with any AI model provider. We have not announced an institutional fundraise. Day-to-day operations are funded by subscription revenue from paying customers, which is the structure that lets us publish honest accuracy numbers and known limitations rather than optimising every claim for a growth narrative.

What is on the roadmap for the next 12 months?

Three themes. First, deeper model coverage as new frontier models ship, with public benchmark updates each quarter so customers can see how detection holds up. Second, workflow surfaces beyond the web app: a stable mobile experience, more integrations into the tools where writers actually work, and a faster streaming API for builders. Third, controls aimed at institutional buyers, including richer team analytics, branded PDF reports, single sign-on, and clearer compliance documentation for academic and enterprise procurement. The roadmap is not a contract, but it is the direction we are committing engineering time to.

Making AI detection accurate & actionable.