Spellcheck can't fix "chai GPT": building an LLM caption checker on $5 a month

Description and use

Misheard leverages an LLM to detect and suggest fixes for misheard words in YouTube auto‑captions, achieving 78 % precision and staying within a $5/month budget.

Problem

YouTube auto‑captions frequently mishear words, producing real‑word errors that standard spellcheckers cannot detect.

Outcome

Misheard improves real‑word error detection from 6/79 to 22/79 on hand‑checked videos, with 0.78 precision, and operates within a $5/month budget.

Workflow

• Upload SRT/VTT file
• Local detectors flag non‑word and phonetic errors
• Chunk transcript (~400 words) and send to LLM with hints
• LLM returns misheard words and suggested fixes
• User reviews and accepts/rejects suggestions
• Export corrected transcript

Evidence

Over half were real‑word errors (55), and only 41 were non‑words.

Source

Read-through: The transcript goes out in chunks of about 400 words. Each chunk carries the local detectors' flags as hints, plus priming terms.

Source

That’s 3.7x and 8.3x the real‑word catches, with better precision, at about $0.05 per hour of audio.

Source