_TRANSCRIBE_PROMPT.md · Stepping App

22 lines · markdown

# Stepping App - Vision Transcription Prompt (for a vision-capable Claude session)
 
Read these two photos and transcribe dad's handwritten list verbatim, then structure it.
 
## Source images (on the homelab, canonical paths)
- `/shared/projects/SteppingApp/IMG_1055.jpg` - the handwritten idea list (bad handwriting; read every line carefully)
- `/shared/projects/SteppingApp/IMG_1056.jpg` - a text-message screenshot (already transcribed, see below)
 
## Already known
- IMG_1056 = text message: "Today 5:19PM - QR code to the stepping app" (dad sending a QR/link)
- Context: this is a "Stepping App" project for dad - likely an app to learn / map out / practise dance "steps" (the handwriting references "promoter classes", "events", "look at videos", "front page", "in each city", "rounds in each"). Treat this as a feature/idea list for the app.
 
## Your job
1. Transcribe IMG_1055 line-by-line, EXACTLY as written. Do not paraphrase. If a word is illegible, write `[?]` next to your best guess.
2. Then produce a clean, de-duplicated feature/idea list grouped by theme (e.g. Content, Events, Location/City, UI/Pages, Monetisation, Promotion).
3. Flag any ideas that seem technically ambitious or unclear so we can clarify with dad.
4. Output as markdown so dad can paste it straight into his planning.
 
## On-box Tesseract draft (unreliable - use only as a hint for hard-to-read lines)
Tesseract could NOT read this handwriting. The only fragments it consistently surfaced:
promoter(s) classes · events · "look at videos" · "should have a live..." · "front page" · "in each city" · "rounds in each" · "free ads/adel" · "3.1 ..." · "Set (eve..." · "NAGS>D ATER"
Do not trust these; the vision read is authoritative.

1	# Stepping App - Vision Transcription Prompt (for a vision-capable Claude session)
2
3	Read these two photos and transcribe dad's handwritten list verbatim, then structure it.
4
5	## Source images (on the homelab, canonical paths)
6	- `/shared/projects/SteppingApp/IMG_1055.jpg` - the handwritten idea list (bad handwriting; read every line carefully)
7	- `/shared/projects/SteppingApp/IMG_1056.jpg` - a text-message screenshot (already transcribed, see below)
8
9	## Already known
10	- IMG_1056 = text message: "Today 5:19PM - QR code to the stepping app" (dad sending a QR/link)
11	- Context: this is a "Stepping App" project for dad - likely an app to learn / map out / practise dance "steps" (the handwriting references "promoter classes", "events", "look at videos", "front page", "in each city", "rounds in each"). Treat this as a feature/idea list for the app.
12
13	## Your job
14	1. Transcribe IMG_1055 line-by-line, EXACTLY as written. Do not paraphrase. If a word is illegible, write `[?]` next to your best guess.
15	2. Then produce a clean, de-duplicated feature/idea list grouped by theme (e.g. Content, Events, Location/City, UI/Pages, Monetisation, Promotion).
16	3. Flag any ideas that seem technically ambitious or unclear so we can clarify with dad.
17	4. Output as markdown so dad can paste it straight into his planning.
18
19	## On-box Tesseract draft (unreliable - use only as a hint for hard-to-read lines)
20	Tesseract could NOT read this handwriting. The only fragments it consistently surfaced:
21	promoter(s) classes · events · "look at videos" · "should have a live..." · "front page" · "in each city" · "rounds in each" · "free ads/adel" · "3.1 ..." · "Set (eve..." · "NAGS>D ATER"
22	Do not trust these; the vision read is authoritative.