Unslop APIv1.3.0

Welcome to Unslop API

Tools for checking your writing and making it sound like you.

Read how Unslop works, no API key needed.

Create an account to get started, or enter an existing API key below. You can check spelling and grammar, find AI-sounding phrases, or rewrite a draft using your own rules.

Prepaid accounts reserve this amount per action and return unused credits. A small cap may stop a multi-call rewrite. Prices and failed-call charges.

Connect to your workspace
Pasted keys stay in this tab's memory. Signed-in accounts use a session cookie.

Sign in or connect with an API key

Text workbench

Analysis runs without AI
0 words20,000 character limit

Editing preferences

Mixed editing sends original excerpts and your guidance to both providers.
English readability estimates, not a guarantee of understanding or suitable subject matter. Ages 6–18 use grade = age minus 5. Use at least 30 words and 2 sentences.Measures text complexity, not the writer's skill. Reading and writing targets must be compatible. Large level changes may need a larger change budget.Added + removed prose tokens ÷ original tokens. A word replacement spends both sides.
More rewrite controlsBy default, AI editing requires at least 50 words and a style score of 20/100. You can change or disable this gate above; the original is always scored. Otherwise your original is returned without a model call. Each call proposes one small change to original text. Unreviewed spans stay unchanged. Meaning review uses one extra call. Paragraph breaks stay intact.Voice comparisons need at least two samples with 20 sentences in total. Paste prose you approve, not instructions.Supports edit count/size, output length, reading level, protected terms and spans. See the API schema below.
Only AI actions send text and selected voice guidance to a model.

API explorer

OpenAPI reference

Run any endpoint, manage private profiles, inspect traps, compare drafts, or connect approved external checks. Profile imports propose rules for you to review before saving.

Request schema

Response

Sign in or connect with an API key, choose an endpoint, and send a request.

A score is a clue.

The style score counts recognizable writing patterns. It does not identify an author. Short passages and quoted examples need context.

Your limits come first.

Code checks edit budgets, protected content, and reading-level constraints. Meaning and voice are still judgment calls; optional AI review is fallible.

Private by default.

Analysis stays on this service. Saved profiles use private workspace storage. External detectors and watermark checks require explicit consent and a configured provider.