Inkbreaker press & media kit

Why we built an engine instead of a bot.

Inkbreaker started from a specific frustration. Writing has always been the craft most resistant to systematization, and the one most damaged by the attempt. Tools that try to "improve" writing by smoothing it toward an average end up teaching writers to sound like each other. That is not craft development. That is drift.

The founder spent over a decade building products at scale across tech, fintech, and Fortune 500 companies. Her most recent role was inside AI-powered educational technology, where she watched the dynamic up close. The conclusion she reached was that writing is one of the domains where AI, as currently implemented, does the most damage to the skill it claims to support.

The damage is subtle. A writer who uses AI to polish a draft stops learning what made the draft rough. A writer who uses AI to generate feedback stops developing the critical eye that feedback is supposed to build. The improvement is borrowed. The skill stays flat.

Inkbreaker is the alternative. Every metric is a count or a ratio computed from the page. Every benchmark is a published standard, calibrated per writing type. Every piece of feedback comes from a human who read the work carefully. Nothing on the platform is predicted by a model or generated for the writer. The engine measures, the community responds, and the writer does the work.

Writers have never had a scored practice environment.

Software engineers have objective practice environments. Platforms where you submit a solution, the system runs it against defined test cases, and you either pass or you do not. The feedback is immediate, specific, and inarguable. No one can tell you your sort algorithm is wrong because of vibes.

Writers have no equivalent. Every writing platform either gives vague encouragement, AI-generated suggestions that blur into the writer’s voice, or academic critique locked behind institutional access. None of them track measurable skill development over time. None of them show you whether your sentence length variation improved between this month and last month.

Inkbreaker is built to fill that gap. The exercise system is modeled on deliberate practice: targeted repetition of specific skills with measurable feedback. The metrics track real linguistic properties: passive voice percentage, syntactic depth, vocabulary range. The skill progression system ties those measurements to a level map so writers can see exactly where they are and what they are building toward.

Inkbreaker is not a gym metaphor. It is a gym. The metrics are the weights, the exercises are the sets, and the report card is the log.

Writing as a measurable skill.

Every craft has a skill map. In software engineering, the map includes time complexity, code readability, test coverage, and system architecture. A developer can measure those properties objectively, track them over time, and practice the weak ones deliberately.

Writing has the same underlying structure but has never had the same infrastructure. Inkbreaker builds it.

The metrics fall into four skill categories:

• Readability: how difficult your prose is to parse. Measured by sentence length, syllable density, and grade-level formulas with decades of research behind them.
• Structure: how you construct sentences and paragraphs. Measured by length variation, average sentence length, and paragraph rhythm.
• Vocabulary: how varied and dense your word choices are. Measured by type-token ratio, lexical density, and repetition score.
• Style: how you handle the elements that distinguish a voice. Measured by passive voice percentage, adverb density, syntactic depth, and sentence opening variety.

These are not grades. They are measurements against benchmarks calibrated for your specific writing type. A fiction writer and a technical writer are not competing. They have different targets because the craft demands are different.

The XP and skill progression system tracks how your metrics move across submissions over time. A writer who consistently improves their Sentence Length Variation score in Fiction exercises earns XP in Pacing. A writer who brings their passive voice percentage down across five Journalism exercises earns XP in Clarity. The connection between measurement and progression is explicit and traceable.

The engine.

Inkbreaker’s prose analysis engine is built on established readability research: Flesch-Kincaid (Kincaid et al., 1975), Gunning Fog (Gunning, 1952), Coleman-Liau (Coleman and Liau, 1975), and silent reading research from Brysbaert (2019). Every metric is a count, a ratio, or a formula applied to those counts.

The engine runs server-side. Submitted text is analyzed and the input is discarded immediately. Genre benchmarks are calibrated per writing type: fiction, poetry, screenwriting, journalism, blogging, technical writing, copywriting, and worldbuilding each have distinct target ranges built from published sources, not from model outputs.

The only place AI touches Inkbreaker is content moderation, where the OpenAI Moderation API is used to flag submissions that violate community guidelines. This is disclosed in the FAQ. Every exercise prompt, every feedback exchange, and every piece of published writing is human-authored.