KeySense AIKeySense AI
v0.4.0 · 70+ layouts · sub-10ms · 94% rule-based

Keyboard Layout Correction API

Fix text typed in the wrong keyboard layout.

KeySense AI detects accidental keyboard layout mistakes — Cyrillic, Arabic, Greek, Hebrew and 66 more — and converts the input into the intended language instantly. No AI required for 94% of corrections.

keysense-ai · live demo
Result will appear here
70+
Keyboard layouts
8
Detection layers
Sub-10ms
Deterministic correction
94%
Rule-based, no AI call

Built for accuracy

Eight layers of detection — from raw glyph mapping to AI verification — with word and phrase-level semantic reconstruction.

Sub-5ms detection

Rule-based pipeline runs entirely on Cloudflare Workers — no external calls, no cold starts. 94% of corrections never touch AI.

🌍

70+ keyboard layouts

Latin, Cyrillic, Arabic, Hebrew, Greek, CJK, and more. Dvorak, Colemak, JIS, ABNT — every major layout is supported.

🧠

Optional AI verification

Low-confidence cases can optionally use AI to improve correction quality. AI is opt-in per-request — most corrections never need it.

🔒

Privacy-first by design

Text data is never stored. API processing is stateless. Enterprise customers can disable external AI entirely for full on-premise operation.

⌨️

Multi-script accuracy

Handles mixed-script inputs and detects the physical keyboard profile (ANSI, ISO, JIS, compact) to maximise correction accuracy.

📦

Official npm SDK

Zero-dependency TypeScript SDK with retry/backoff, AbortSignal, and full type coverage. One install away.

One npm install away

TypeScript
import { KeySenseClient } from '@keysense/sdk';

const ks = new KeySenseClient({
  apiKey: process.env.KEYSENSE_API_KEY,
});

const result = await ks.detect('руддщ цщкди');
// → { decoded: 'hello world', layout: 'ru-RU', confidence: 1.0 }

const result = await ks.detect('ηεββο ςορβδ', { useAI: true });
// → { decoded: 'hello world', layout: 'el-GR', confidence: 0.95 }
npm install @keysense/sdk

Built for real workflows

Whether you're cleaning CRM data, fixing search queries, or building a multilingual product — KeySense AI slots in as a single API call.

🔍

Search boxes

Fix mis-typed queries before they hit your search engine. Users who accidentally type in the wrong language still get the right results.

💬

Customer support

Auto-correct incoming ticket text and chat messages. Agents spend less time deciphering garbled input.

🌐

Multilingual apps

Your product serves users across many countries — KeySense AI catches layout errors before they surface as bugs or bad UX.

🗂️

CRM / form input

Names, addresses, and notes typed with the wrong layout are silently corrected before they pollute your database.

✍️

Transliteration cleanup

Convert phonetically-typed Cyrillic, Arabic or Greek back to the intended native script, or vice versa.

🏢

Enterprise on-premise

Deploy the worker in your own Cloudflare account or private cloud. No data ever leaves your infrastructure.

AI is optional

Your data never leaves unless you say so

94% of corrections are handled entirely by our rule-based engine — no AI, no external calls, no data sharing. The Claude AI fallback is opt-in and can be disabled per-request or globally. Enterprise customers can deploy on-premise and disable external AI processing entirely.

Read the AI fallback docs →

Simple pricing

Start free. Scale when you need to.

Free
$0/month

For individuals and small projects.

  • 500 corrections / month
  • 70+ layouts
  • Rule-based detection only
  • @keysense/sdk access
  • Community support
Get started free
Most popular
Pro
$9/month

For teams and production apps.

  • 50,000 corrections / month
  • AI verification included
  • Priority API access
  • Usage dashboard
  • Email support
Start Pro trial
Enterprise
Custom

For high-volume or on-premise needs.

  • Unlimited corrections
  • Dedicated worker instance
  • On-premise deployment
  • SLA guarantee
  • Dedicated support
Contact us
FeatureFreeProEnterprise
Corrections / month50050,000Unlimited
API access
TypeScript SDK
70+ keyboard layouts
Rule-based engine
AI verificationopt-in
Sensitive modedisables external AI per-request
Disable external AI globally
On-premise deployment
Usage dashboard
Rate limit60/min600/minCustom
Log retention30 days30 daysCustom
SupportCommunityEmailDedicated SLA
DPA / GDPR docs

Need a custom volume deal or private deployment? Contact us for Enterprise