IELTS Speaking Coach

An OpenClaw skill that acts as a professional IELTS Speaking Examiner and Tutor. Get instant, evidence-based assessment of your spoken English with actionable feedback — including real pronunciation scoring from audio input.

Features

• Four-criterion scoring (Band 1-9, 0.5 increments): Fluency & Coherence, Lexical Resource, Grammar, Pronunciation — each with specific transcript evidence
• Audio pronunciation scoring: Send voice messages for real PR band scores via ffmpeg + ASR (text-only mode estimates PR from other criteria)
• CHAI-calibrated scores: Human-prior calibration adjusts for Part difficulty and criterion bias
• Mock exam simulation (Menu 7): Full Part 1→2→3 timed test with no in-test feedback and comprehensive final report
• ZPD learning paths (Menu 6): Personalized vocabulary and grammar progression based on your current band level
• Adaptive difficulty: Questions automatically adjust to your band level
• Grammar corrections: Identifies genuine spoken grammar errors with minimal corrections and rule explanations
• Context-aware vocabulary upgrades: Topic-specific collocational improvements, not generic thesaurus swaps
• Pronunciation guidance: Targeted tips for Chinese speakers covering th/v/r/l, stress, rhythm, intonation, and connected speech
• Spoken-register model answers: Rewrites at your target band using natural discourse markers, contractions, and idiomatic language
• All three Parts: Part 1 (Interview), Part 2 (Long Turn), Part 3 (Discussion) with Part-specific expectations
• Menu-driven practice mode: Fixed entry phrases, numbered menu options, and standardized output templates
• 42 difficulty-tagged cue cards: Part 2 practice with [Easy]/[Medium]/[Hard] cards across 7 topic categories

Installation

openclaw skills install ielts-speaking-coach

Quick Start

Assess a response (text)

Score my IELTS Part 1 answer: "I'm studying computer science at university. It's quite challenging but I find it rewarding because I enjoy problem-solving."

The skill returns a full assessment: scores, grammar corrections, vocabulary upgrades, model answer, and ZPD learning recommendations.

Assess with audio

Send a voice message and the skill will:

1. Convert audio via ffmpeg
2. Transcribe using ASR
3. Score pronunciation based on ASR confidence
4. Provide full four-criterion feedback with real PR scores

Practice session

Let's do IELTS speaking practice, Part 2

The skill gives you a difficulty-appropriate cue card, listens to your response, and provides detailed feedback.

Mock exam

模拟考试

Full 11-14 minute IELTS simulation: Part 1 (4-5 min) → Part 2 (3-4 min) → Part 3 (4-5 min), followed by a comprehensive score report.

Learning path

学习路线

Get a personalized ZPD-based vocabulary and grammar study plan based on your current band level.

Configuration

Option	Type	Default	Description
target_band	number	7.0	Your target IELTS band (model answers aim at this level)
default_part	number	1	Default IELTS Speaking part (1, 2, or 3)
user_id	string	"default"	Learner ID for trajectory tracking

openclaw skills config ielts-speaking-coach target_band 7.5
openclaw skills config ielts-speaking-coach default_part 2

Fixed Entry Mode

Use one of these phrases to enter the guided practice flow:

• 进入雅思口语模式
• 启动雅思口语教练
• Use IELTS speaking coach

The skill then shows a fixed menu:

1. Part 1 练习
2. Part 2 练习
3. Part 3 练习
4. 口语评分
5. 语法纠错
6. 学习路线
7. 模拟考试

You can reply with a number or paste a transcript directly.

How Scoring Works

The skill evaluates four criteria independently using official IELTS Band Descriptors:

1. Fluency & Coherence: Speech rate (WPM), filler density, discourse markers, coherence
2. Lexical Resource: Vocabulary sophistication, semantic variety, idiomatic usage, collocations
3. Grammatical Range & Accuracy: Complex structures per 18 words, error density
4. Pronunciation: ASR confidence-based scoring for audio input; median estimation for text-only

Scores are calibrated using the CHAI framework (hybrid human-prior calibration) with Part-specific adjustments, then averaged and rounded to the nearest 0.5 band.

Permissions

Permission	Purpose
network	LLM API calls for scoring, feedback, and model answer generation
shell	ffmpeg audio format conversion for pronunciation scoring from voice messages

Bundled Resources

• scoring-rubric.md — Official IELTS Band descriptors
• cue-cards-2025-may-aug.md — 15 official cue cards (2025 May–August bank)
• cue-cards.md — 42 difficulty-tagged cue cards across 7 categories
• learning-path.md — ZPD vocabulary/grammar progression reference
• pronunciation-guide.md — Chinese speaker pronunciation diagnostics and exercises
• vocab-map.json — Topic-aware vocabulary upgrade hints
• examples.md — Sample interactions

Self-hosted Backend (Optional)

The GitHub repository includes an optional backend/ directory with a FastAPI server that provides enhanced features:

• DL model-based precise scoring
• Persistent learning state tracking across sessions
• Vocabulary ontology-driven trajectory planning

The backend is not required for the skill to work. All core features (scoring, practice, mock exam, learning paths) run entirely via the built-in LLM and bundled reference files. See the GitHub repository for backend setup instructions.

License

MIT