skills.rest
HuuBar
HuuBarCOMMUNITY
Explore all skills by this author →
GitHub

transcript-fixer

Correct ASR transcripts using dictionary rules and AI-assisted edits.

Updated Apr 21, 2026
Search Tags:#workflow#transcript#learning#dictionary#asr#ai-correction
One-click install
npx skills add https://github.com/HuuBar/skill-routing-experiment --skill transcript-fixer-huubar
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: transcript-fixer
Source: https://github.com/HuuBar/skill-routing-experiment/tree/main/unified_skills/daymade/transcript-fixer
Command: npx skills add https://github.com/HuuBar/skill-routing-experiment --skill transcript-fixer-huubar

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires httpx, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Transcripts produced by speech-to-text systems often contain recognition errors and misinterpretations that slow down review and dissemination. Transcript Fixer combines deterministic dictionary rules with AI-powered corrections to deliver a clean, accurate transcript.

Core Features & Use Cases

  • Two-phase correction pipeline: deterministic dictionary rules for instant fixes, followed by AI-powered corrections for remaining errors.
  • Learns from each fix and stores improvements in ~/.transcript-fixer/corrections.db to build a personalized correction dictionary over time.
  • Handles ASR outputs with recognition errors, homophones, mixed Chinese/English terms, and garbled jargon; can also clean meeting notes and lecture transcripts.
  • Supports batch processing, domain-specific dictionaries, and diff reporting to track improvements.

Quick Start

Feed a transcript into the tool to obtain a corrected, ready-to-review version.

Frequently Asked Questions about transcript-fixer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix ASR transcript errors automatically?▼

To fix ASR transcript errors automatically, use a two-phase pipeline that applies deterministic dictionary rules for instant corrections followed by AI-powered edits for remaining recognition errors to produce a clean transcript.

What is the best way to correct mixed Chinese and English terms in speech-to-text outputs?▼

Correcting mixed Chinese and English terms in speech-to-text outputs is handled through AI-powered corrections and domain-specific dictionaries, which resolve homophones and garbled jargon effectively.

Can I use a personalized correction dictionary for meeting notes and lecture transcripts?▼

Yes, you can use a personalized correction dictionary for meeting notes and lecture transcripts. The system learns from each fix and stores improvements locally to enhance future corrections.

How does diff reporting work when processing ASR corrections?▼

Diff reporting for ASR corrections works by tracking the changes made during the dictionary-based and AI-assisted editing phases, providing a clear comparison to review improvements and verify domain-specific safety checks.

Do I need httpx to run batch processing for transcript fixes?▼

Yes, you need httpx installed to run batch processing for transcript fixes, as it is the required dependency for handling the AI-powered corrections and external requests within the workflow.

Related Skills

Explore More Software Engineering Skills→
daymadedaymade
1.3k

transcript-fixer

Correct ASR transcription errors using dictionary rules, GLM AI, and SQLite learning.

Community
Advanced
liangdabiaoliangdabiao
71

humanizing-expert

Rewrite AI-generated text into natural, human-like writing.

Community
Basic
Team-HirealTeam-Hireal

openai-whisper

Transcribe and translate local audio files using the Whisper CLI.

Community
Basic
a554b554a554b554
53

proofread

Correct grammar, spelling, and punctuation errors in text while respecting scope modifiers.

Community
Basic