Your life. Your voice. Your book.
Transform boxes of physical artifacts — letters, journals, photos, negatives, home videos — into searchable digital archives, memoir scaffolds, and AI writing companions.
Named after T.S. Eliot's The Love Song of J. Alfred Prufrock.
- Intake — Inventory, categorize, and process incoming materials
- Transcribe — OCR handwritten pages using Claude Vision (letters, margins, journals)
- Faces — Detect and cluster faces across all photos for identification
- Timeline — Extract dates from transcriptions and EXIF data, build chronology
- Assemble — Merge everything into a client-ready deliverable package
pip install -e .For facial recognition:
pip install face-recognition # requires dlib# 1. Process incoming materials
prufrock intake ./client-box/
# 2. Transcribe handwritten pages
export ANTHROPIC_API_KEY=sk-...
prufrock transcribe ./client-box/scanned-pages/ \
--context "Family: John (father), Pat (mother), Rob, Thomas, Bill, Kathleen. Location: Warwick, NY."
# 3. Cluster faces across photos
prufrock faces ./client-box/photos/
# 4. Build timeline from all sources
prufrock timeline ./client-box/
# 5. Assemble final deliverable
prufrock assemble ./client-box/ --client-name "The Chuvala Family"Physical Box
↓
[intake] → Inventory, resize, categorize
↓
[transcribe] → Claude Vision OCR → tagged markdown
↓
[faces] → DeepFace clustering → ID worksheet
↓
[timeline] → Date extraction → chronology
↓
[assemble] → Merged deliverable package
↓
Client Archive + Memoir Scaffold
| Tier | Deliverable |
|---|---|
| Archive | Digitized, tagged, cataloged, searchable, timeline |
| Story | Archive + thread analysis + memoir scaffold + walkthrough |
| Companion | Story + custom AI spec with client voice profile |
- Python 3.10+
- Anthropic API key (for transcription)
- dlib + face_recognition (for facial clustering)
MIT
"Do I dare disturb the universe?"
Part of the Northwoods Sentinel Labs ecosystem — open-source tools for human-centered AI.