tomevault-io EPUB to Markdown
When to Use
User wants to convert EPUB file(s) to Markdown.
Prerequisites
Run ./scripts/check-prereqs.sh. If "ready": false:
| Missing | Action |
|---|---|
| node | Direct user to install Node.js 18+ from nodejs.org |
| epub2md | Offer to run npm install -g epub2md (requires user confirmation) |
If epub2md is the only missing dependency, ask: "epub2md is not installed. Install it now?" If yes, run the install command and re-run check-prereqs.sh to confirm.
Workflow
1. Validate Input
ls -la "<epub_path>" # or: ls <wildcard_pattern>
2. Create Output Directory
Derive from EPUB filename to avoid overwrites:
output_dir="${epub_path%.epub}_md"
[ -d "$output_dir" ] && output_dir="${output_dir}_$(date +%s)"
mkdir -p "$output_dir"
3. Convert
cd "$output_dir" && epub2md "../<epub_file>"
Options (ask user preference if not specified):
--merge- Single file instead of per-chapter files--localize- Download remote images locally (Node 18+)-M- Fix Chinese/English spacing
Batch:
for f in <pattern>; do d="${f%.epub}_md"; mkdir -p "$d"; (cd "$d" && epub2md "$f" --merge); done
4. Report
Show file count and location. If issues: epub2md -S <file> shows structure, -i shows metadata.
Errors
| Error | Cause |
|---|---|
| "Cannot find module" | epub2md not installed globally |
| Empty output | DRM-protected or corrupted EPUB |
Feedback
Hybrid: script + qualitative.
1. Run Evaluator
./scripts/evaluate.sh "<epub_path>" "<output_dir>"
Checks size sanity (output ≤ input) and structure (has .md files).
2. Ask User
"Does the Markdown preserve the content you expected?"
Map: great→5, good→4, minor issues→3, problems→2, broken→1
3. Log
Append to FEEDBACK.jsonl:
{"ts":"<ISO8601>","skill":"epub-to-md","version":"0.1.0","prompt":"<request>","outcome":<1-5>,"note":"","source":"hybrid","schema_version":1}
Increment iteration_count in CONFIG.yaml.
Converted and distributed by TomeVault — claim your Tome and manage your conversions.