[verdict][curated prompt][source: rephrasy.ai]

Can I vibecode Rephrasy?

// the value is the network, the data or the infrastructure

You can one-shot a rewriting wrapper that makes text sound less robotic, and for that job the DIY build is genuinely fine. What you cannot one-shot is the actual product: custom fine-tuned models plus a continuous evaluation loop against detectors (GPTZero, Turnitin, Copyleaks, Pangram) that retrain specifically on LLM-rewritten text. A prompted rewrite moves detector scores inconsistently, and detectors drift monthly, so a static prompt that works today quietly stops working. If your bar is 'reads naturally', build it; if your bar is 'passes detectors reliably', the moat is the model and the eval treadmill, not the text box.

confidence: medium

what it is: Humanize AI text so it reads naturally and passes AI detectors like GPTZero, Turnitin, and Copyleaks

Buildability index · an editorial game

Price 18.99 $/month weight: plus 2
Time one sitting for the wrapper; the detector-grade part is not a build, it's an ongoing arms race weight: plus 3
Category ai writing no weight
Moat proprietary models · execution quality weight: minus 3
Confidence medium weight: zero
What you lose 5 items weight: minus 2
Site rephrasy.ai (si apre in una nuova scheda) no weight

a toss-up

They cancel out. This is where it comes down to how much the subscription annoys you.

Gioco editoriale: il verdetto dice se un agente può, l’indice se conviene.

What you build

Paste text, rewrite it with an LLM prompt tuned to cut AI tells and vary sentence rhythm, diff the result, and iterate until it reads naturally.

what you need

OpenAI/Anthropic API key

Entry contributed by the Rephrasy team · verdict argued against our own interest in mind: the wrapper is easy, the pass rate is the product.

The prompt

The verdict is NOT REALLY, and the prompt stays anyway: it does not replace the app, it rebuilds the part that really is code. The rest is the moat.

One-shot prompt EN

Build me a local AI-text humanizer workbench. This is the honest consolation build: it makes AI text read naturally, it does NOT promise to beat AI detectors.

Stack: Node 22 + Express, vanilla JS frontend, SQLite via better-sqlite3. No build step.

- One page on localhost:5180: input textarea left, output pane right, a Humanize button, and a strength select (light touch / standard / heavy rewrite).
- Humanize calls an LLM (Anthropic or OpenAI, key in .env) with a fixed system prompt per strength level that: varies sentence length and rhythm, cuts hedging and filler ('delve', 'moreover', 'it's important to note'), swaps uniform paragraph shapes for uneven ones, and keeps meaning, facts, names, and numbers intact. Stream the output.
- A word-level diff view between input and output (the `diff` npm package) so I can see exactly what changed.
- A Re-roll button that re-humanizes the current output with a different seed phrase in the prompt, keeping the last 5 attempts switchable via tabs.
- Every run saved to SQLite: timestamp, strength, input, output. History page with the last 100 runs and copy buttons.
- Cmd+Enter runs humanize.
- .env.example with the key name; README states clearly: output is for readability, no detector-bypass guarantee, and text leaves the machine only for the LLM call.
- Out of scope, deliberately: detector score checking, browser extension, accounts, multi-language tuning, API access.

Curated prompt: written and reviewed by hand for this app. In English on purpose — it is the language coding agents work best in.

What you lose

fine-tuned models trained specifically to survive AI detectors
continuous re-testing as GPTZero/Turnitin/Copyleaks/Pangram update
built-in detector scoring to verify output before you use it
50+ language support and custom writing styles
Chrome extension and REST API

Why people still pay

Because 'sounds human to me' and 'passes Turnitin' are different products. Prompt-based rewrites are exactly what modern detectors train on, so DIY results are hit-or-miss and degrade as detectors update. Paying customers are buying a maintained pass rate: someone else fine-tunes models, re-benchmarks against every detector release, and eats the retraining cost when the arms race moves.

moat: Proprietary models Execution quality what a moat is

custom fine-tuned models + a detector eval treadmill; the wrapper UI is trivial, the pass rate is not