Wordferry

Edited English samples that help Indian-language books sell translation rights

SMEs

TL;DR Wordferry turns the first chapters of a Kannada, Marathi or Bengali book into a polished English sample, with a synopsis and pitch pack, so the book can be shown to English publishers and agents. A machine does the first draft and a literary editor does the real work. Regional publishers and authors pay per title.

The problem

Most good books written in Indian languages never reach an English editor's desk, because nobody wants to pay for a full translation before a publisher is even interested. When one does get through, the upside is real - Banu Mushtaq's Kannada stories, translated by Deepa Bhasthi as Heart Lamp, won the 2025 International Booker, the first Kannada work to do so[1].

So can't the cheap tools fill the gap? Not yet. One giant online retailer's e-book translation tool, launched in beta in November 2025, covered only English, Spanish and German, and the retailer itself warned that its output can contain errors[2]. A raw machine draft of a novel is not something an editor at a big house will read past page three.

The idea

Wordferry sells a rights-ready sample, not a full translation. For each book it produces about 8,000–12,000 words of English (usually the opening chapters), a two-page synopsis, an author note and a one-page market comparison. An Indian-language AI model, such as the ones Indian AI startups now offer[3], makes the first pass. A paid literary translator who knows the source language then rewrites it for voice, idiom and rhythm, and a second editor checks it against the original.

If a publisher bites then the same translator is offered the full book on a separate contract, and Wordferry takes a placement fee.

The seed pitched a translation platform for self-published authors, which I don't buy, because an indie author rarely earns enough from English sales to pay for a full literary translation. The sample is the cheap, testable step that tells everyone whether the full job is worth paying for.

Regulatory note: no licence is needed, but the author must hold, and grant in writing, the translation right in the work[4].

Who pays

There are two buyers. Small and mid-sized regional-language publishers own backlists and want English rights income without hiring translators on spec. Established regional authors with prizes or strong local sales will pay for their own sample. There's no public rate card for Indian literary translators, so test pricing directly. My working guess is ₹40,000–90,000 per sample package, plus a success fee on any rights deal.

Why now

The 2025 Booker win put Indian-language fiction in front of English editors and readers[1]. Indian-language AI models now handle the dull first draft in minutes[3], which cuts the human hours per sample sharply. And the big retailer's tool skipped Indian languages at launch[2], so the field is open for now.

Market signal

There is no public figure for spending on translation samples or rights pitching in India. The signal is qualitative - a prize-winning translated collection[1] and a large AI firm building Indian-language translation and narration tools[3]. Treat the Indian publishing industry as a whole, put at about ₹781 billion for 2024 in one projection cited by a publishing-services firm, as adjacent context only[5].

Who's already doing this

The retailer's translation tool is the closest automated tool, but it handled no Indian languages at launch[2]. One Indian AI firm sells machine translation and narration, not edited literary samples[3]. Freelance translators and a few literary agents do this work one book at a time, often unpaid and on spec. The gap is a repeatable service with fixed prices and a steady pipeline of editors.

How hard is it?

Difficulty Moderate (31/100)
Starting capital ₹30–55 L
Time to launch 4–6 months

The software is the easy part. The hard part is finding and keeping literary translators who are good in both languages and will accept a machine draft as a starting point, which is a bigger ask than it sounds. Publishers' trust comes next, because one weak sample sent under your name to an English editor costs you that editor.

Licences and approvals

No sector licence is needed to run this. The real permission is the translation right, which you get from the author or publisher, not from a regulator.

  • Written grant of the translation right from the author or rights holder, before you translate a single page. Under the Copyright Act, 1957, translation is the author's exclusive right[4].
  • Company or LLP registration with the MCA before you sign your first publisher.
  • GST registration once turnover crosses the threshold, since your fees are taxable services.
  • DPDP Act compliance before launch, because you hold author and translator contact data[6].

Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.

How AI can power this business

  • First-draft translation. An Indian-language model drafts the sample chapters, so the literary translator starts from a rough English text and spends their hours on voice and rhythm[3].
  • Consistency checks. A model flags names, places and recurring terms that drift between chapters, which saves the second editor a slow read.
  • Pitch pack drafts. It can draft the synopsis and comparable-titles list from the manuscript for the editor to rewrite, cutting a day off each package.
  • Pipeline triage. Summaries of incoming manuscripts help you decide which books are worth sampling at all.

Don't trust AI with the final literary text or with judging whether a book is good. That's the editor's job, and one weak sample costs you an English editor's attention.

Want to work out the AI layer for Wordferry? Talk to Yash →

Biggest risks

  • Translator supply. Good literary translators are few, and many dislike editing machine output.
  • Slow payoff. A rights deal can take a year or more, so success fees arrive late and unpredictably.
  • Quality reputation. If editors see Wordferry samples as machine work then every later pitch is discounted.

Our take

I think this suits a founder from publishing or literary translation who already knows editors at English houses. For a pure technologist it's a weak fit. Before raising money, prove that 10 regional publishers will each pay for one sample, and that at least two of those samples get a full-manuscript request within six months.

Questions founders ask

Plan for roughly ₹30–55 L, most of it going on translator and editor fees while you build a pipeline of titles.

Do I need a licence to start a translation service in India?

No sector licence. You need a written grant of the translation right from the author or publisher, plus company registration, GST once over the threshold and DPDP compliance.

Can I start a translation rights service in another country?

Yes, if you know a literary market with strong books in a language English publishers rarely read. Copyright law on translation rights differs by country, so take local legal advice first.

Download the kit

The free Wordferry validation kit has the line-by-line capital breakdown, the effort scoring, the rights and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for publishers and authors, and 5 assumptions to test. Download the kit →

Talk to Yash

Thinking about building Wordferry or something close to it? The make-or-break work is building a bench of literary translators and getting English editors to read your first samples, and that's the bit I like helping founders with. If you need a hand, you know where to find me. Talk to Yash →

Sources

  1. Wikipedia, "Heart Lamp", accessed 2026-10-21. https://en.wikipedia.org/wiki/Heart_Lamp
  2. LatestLY, "Amazon introduces Kindle Translate, AI-powered service for KDP authors", 2025-11-07. https://www.latestly.com/technology/amazon-introduces-kindle-translate-ai-powered-service-will-boost-kindle-direct-publishing-authors-to-reach-global-readers-7195572.html
  3. Sarvam, "Text to speech for audiobooks" (official site), accessed 2026-10-21. https://www.sarvam.ai/text-to-speech/audiobooks
  4. PRS Legislative Research, "The Copyright (Amendment) Bill, 2010", accessed 2026-10-21. https://prsindia.org/billtrack/the-copyright-amendment-bill-2010
  5. PublishDrive, "New store available: Repro India", 2023-08-10. https://publishdrive.com/new-store-available-repro-india.html
  6. PRS Legislative Research, "Digital Personal Data Protection Bill 2023", 2023-08. https://prsindia.org/billtrack/digital-personal-data-protection-bill-2023

For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.

Reviews

No reviews yet