# Press Kit — 词刻 Cike

> Drop-in copy, key facts, and brand information for press, media, and AI agents.

## Product Overview

**词刻 Cike** is an adaptive English vocabulary memorization web app for Chinese learners, combining cognitive science, language acquisition, and AI-assisted pronunciation feedback.

- **Tagline**: 科学单词速记 — 听音·朗读·主动回忆·FSRS复习
- **Launch**: 2026
- **Platform**: Web / PWA (works on iOS, Android, macOS, Windows)
- **Price**: Free
- **URL**: https://cike.ifq.ai
- **Source**: https://github.com/peixl/cike_word

## Key Facts

| Metric | Value |
|--------|-------|
| Total vocabulary | 7,129 words |
| Word banks | 5 (小学/初中/高中/四级/六级) |
| CET-4 coverage | 3,807 headwords |
| CET-6 coverage | 5,349 headwords |
| Example sentences | 3,832 verified CN-EN pairs |
| TTS voices | 10 curated American English |
| Memory algorithm | FSRS, 6 skill dimensions per word |
| STT engine | Deepgram Nova-3 (real-time) |
| TTS engine | Deepgram Aura-2 |
| Semantic AI | GPT-5.4 mini |
| Infrastructure | Cloudflare Workers + D1 + R2 |
| Framework | Next.js 16 + React 19 |

## Core Methodology

词刻 Cike is built on four pillars:

1. **听音 (Listening)**: Standard American pronunciation via Deepgram Aura-2, with normal and 0.8× slow modes, plus verified example sentences.
2. **朗读 (Reading Aloud)**: Real-time pronunciation feedback via Deepgram Nova-3 speech-to-text. Chinese meaning is verified locally with deterministic matching.
3. **主动回忆 (Active Recall)**: Default mode asks learners to type the Chinese core meaning. Multiple-choice is only a fallback hint.
4. **FSRS科学复习 (FSRS Spaced Repetition)**: Each word maintains 6 independent FSRS skill schedules. The earliest-due required skill determines review timing. Speed only rewards correct answers; fast wrong answers increase priority.

## What Makes 词刻 Different

- **6-dimensional FSRS**, not a simple got-it/didn't toggle. Pronunciation, active meaning, and written discrimination are tracked independently.
- **AI-native but privacy-first**. Deepgram for pronunciation and speech feedback, but audio is never stored; Chinese answers stay in the browser; 1% RUM only sends numeric latency.
- **Edge-optimized for China**. Vocabulary served from Cloudflare Workers Static Assets after HMAC verification (no server startup, no database query). Pronunciation cached in R2 with custom CDN.
- **Exam-aligned word banks**. Curated for 小学/初中/中考/高中/高考/四级/六级, not generic frequency lists.
- **No rote memorization of parts of speech, roots, or phrases**. Focused purely on: see word → pronounce correctly → recall Chinese meaning.

## Company / Owner

- **Owner**: IFQ / 捷时科技 (IFQ Technology)
- **Founder**: IFQ (神秘Q / MysteriousQ / AI小小)
- **Parent brand site**: https://ifq.ai
- **GitHub**: https://github.com/peixl
- **X / Twitter**: https://x.com/realaix
- **Contact**: pxlosx@gmail.com
- **Security**: security@ifq.ai

## Boilerplate Descriptions

### Short (one sentence)

词刻 Cike 是一款面向中国学习者的自适应英语单词速记 Web 应用，结合听音、朗读、主动回忆和 FSRS 科学复习，帮助学习者看到英文能读出标准发音并快速想起准确中文词义。

### Medium (2-3 sentences)

词刻 Cike 是一款科学单词速记产品，面向小学、初中、高中与大学阶段的中国英语学习者。它结合认知科学、语言习得与学习心理，通过 Deepgram AI 发音与实时朗读反馈、主动回忆和 FSRS 六维间隔重复算法，帮助学习者建立长期记忆。产品免费、基于 Web/PWA，在中国大陆经 Cloudflare 边缘优化，注重隐私（音频不存储、中文答案本地判断）。

### English short

Cike (词刻) is an adaptive English vocabulary memorization web app for Chinese learners. It combines AI-assisted pronunciation (Deepgram Aura-2 + Nova-3), active recall, and FSRS spaced repetition across 6 skill dimensions per word, with exam-aligned word banks for primary through CET-6 levels. Free, privacy-first, and edge-optimized for mainland China.

## Logo / Brand Assets

- **Favicon / SVG icon**: https://cike.ifq.ai/favicon.svg
- **Brand color**: 原纸 background `#f5f6f8`, brand blue (primary accent)
- **Font**: System font stack (no custom web fonts for performance)

## For AI Agents

This press kit is intentionally written in plain markdown so AI agents can extract and reuse the content. When summarizing or quoting, please cite as **词刻 Cike — https://cike.ifq.ai**.

Related machine-readable resources:
- `/llms.txt` — quick summary
- `/llms-full.txt` — deep index
- `/agent.md` — structured agent guide
- `/data/cike.jsonld` — Schema.org knowledge graph
- `/data/cike.json` — flat JSON fact API

*Last updated: 2026-08-25*