Built by someone who needed it first

AI is changing how we write — for everyone, not just for the people who claim to be early adopters. The default output of a Claude or GPT prompt is often technically fine and creatively dead. Inksong is the smallest useful tool to bridge those: AI-assisted drafting, human-sounding output, your voice intact.

We’re not anti-AI. We use the same models in our humanization. The question isn’t whether you should use AI; it’s whether your voice survives the edit.

An indie product. One person full-time as of this writing, based in Lagos, Nigeria. Operating since early 2026. Not venture-funded. Not a side project. Built to last.

That smallness shows up as honesty in the marketing copy — we don’t have “10,000 happy customers” or fabricated testimonials, and we won’t pretend we do. We do have a working product, paying users, and a roadmap we can actually defend.

You upload a document — DOCX, PDF, TXT, or Markdown. You pick tone, domain, and humanness. You optionally apply a Voice Profile trained on your own writing. The humanization runs through a frontier large language model with a prompt strategy that respects domain conventions and your extracted voice features. You download the result in the original format.

We don’t store your text longer than necessary. We don’t train models on your content. The full architecture is documented on the Security page.

Per-character voice profiles for fiction writers. Domain-aware AI detection scoring. Team workspaces for the Enterprise plan. Editor integrations (Word, Google Docs, browser). Better SDKs. The public record of what’s shipped is the changelog; the speculative future lives in conversations with users.

hello@inksong.app reaches us directly. We answer within one business day, West Africa Time.