Curator

3. Background Agent

Long-running autonomous acquisition:

Task(subagent_type="curator", prompt="Continuous acquisition: build Paracelsus collection", run_in_background=true)

Thematic Focus

Primary Collections

Secondary Collections

• Early Science - Copernicus, Kepler, Newton
• Neoplatonism - Plotinus, Proclus, Florentine Academy
• Emblemata - Alciato, Atalanta Fugiens, Splendor Solis
• Architecture - Vitruvius, Palladio, sacred geometry
• Esoteric Music - Pythagorean harmony, musica universalis, cosmic music theory

Esoteric Music Focus

Languages (priority order)

1. Latin (primary scholarly language)
2. German (Boehme, Paracelsus)
3. English (17th century translations)
4. Italian (Renaissance sources)
5. French (18th century editions)
6. Dutch (Amsterdam printing)

Selection Rules (CRITICAL)

Edition Priority (CRITICAL)

ALWAYS prefer the oldest available edition in original language:

1. Incunabula (pre-1501) - Highest priority
2. 16th century editions - First printed editions, editio princeps
3. 17th century editions - Important scholarly editions
4. 18th century editions - When earlier unavailable
5. 19th century critical editions - Scholarly Latin/Greek texts (e.g., Teubner, Loeb)
6. Modern translations - ONLY when no original text edition exists

Language Priority:

• Original language (Latin, Greek, German, Arabic) ALWAYS over English translations
• Contemporary translations (e.g., 17th c. English of Latin) acceptable as supplements
• NEVER import 20th-21st century English translations when Latin/Greek originals exist

ACQUIRE

• Original historical editions (pre-1800 primary sources)
• Early printed books in original language
• First editions and important early printings
• 16th-17th century Greek/Latin scholarly editions
• Contemporary translations (17th-century English of Latin works)
• Critical scholarly editions with original text (Teubner, etc.)

REJECT

• Modern translations (20th-21st century) without original text
• English-only editions when Latin/Greek available
• Secondary literature and commentaries
• Facsimile reprints when original scans exist
• Anthologies that excerpt rather than present complete works
• Books already in collection (check before importing)

Scoring (1-10 scale)

API Reference

Check Existing Collection

# Search for author/title to avoid duplicates
curl -s "https://sourcelibrary.org/api/search?q=AUTHOR_OR_TITLE&limit=20"

# Get collection stats
curl -s "https://sourcelibrary.org/api/admin/stats"

# List all books
curl -s "https://sourcelibrary.org/api/books" | jq '[.[] | {id, title, author, year}]'

Import from Internet Archive

curl -s -X POST "https://sourcelibrary.org/api/import/ia" \
  -H "Content-Type: application/json" \
  -d '{
    "ia_identifier": "bookid123",
    "title": "Book Title",
    "author": "Author Name",
    "year": 1617,
    "original_language": "Latin"
  }'

Import from Gallica (BnF)

curl -s -X POST "https://sourcelibrary.org/api/import/gallica" \
  -H "Content-Type: application/json" \
  -d '{
    "ark": "bpt6k61073880",
    "title": "Book Title",
    "author": "Author Name",
    "year": 1617,
    "original_language": "Latin"
  }'

Import from MDZ (Bavarian State Library)

curl -s -X POST "https://sourcelibrary.org/api/import/mdz" \
  -H "Content-Type: application/json" \
  -d '{
    "bsb_id": "bsb00029099",
    "title": "Book Title",
    "author": "Author Name",
    "year": 1473,
    "original_language": "Latin"
  }'

Acquisition Workflow

Phase 1: Discovery

# Search Archive.org by author
curl -s "https://archive.org/advancedsearch.php?q=creator:(Paracelsus)+mediatype:(texts)+date:[1500+TO+1700]&output=json&rows=50" | jq '.response.docs[] | {identifier, title, date, creator}'

# Search Gallica
# Use web search for: site:gallica.bnf.fr "Author Name"

# Search MDZ
# Use web search for: site:digitale-sammlungen.de "Author Name"

Phase 2: Evaluation

For each candidate:

1. Check if already in collection (search API)
2. Verify it's a primary source, not modern translation
3. Score against criteria (aim for 7+/10)
4. Note edition details, page count, image quality

Phase 3: Import

# Import and capture book ID
RESP=$(curl -s -X POST "https://sourcelibrary.org/api/import/ia" \
  -H "Content-Type: application/json" \
  -d '{"ia_identifier": "...", "title": "...", "author": "...", "year": ...}')

BOOK_ID=$(echo "$RESP" | jq -r '.book.id // .id')
echo "Imported: $BOOK_ID"

Phase 4: Queue Processing

After import, queue for OCR:

# Create batch OCR job
curl -s -X POST "https://sourcelibrary.org/api/jobs" \
  -H "Content-Type: application/json" \
  -d "{
    \"type\": \"batch_ocr\",
    \"book_id\": \"$BOOK_ID\",
    \"model\": \"gemini-2.5-flash\",
    \"language\": \"Latin\"
  }"

Reporting Format

Per-Book Report

## [Title] ([Year])
**Author**: [Name]
**Language**: [Lang] | **Pages**: [N] | **Source**: [archive.org ID]
**Theme**: [Primary collection]
**Score**: [N]/10
**Notes**: [1-2 sentences on significance]
**Status**: [acquired/processing/complete]

Batch Summary

# Acquisition Batch [DATE] - [THEME]

## Summary
- Books acquired: N
- Total pages: N
- Languages: X, Y, Z
- Date range: YYYY-YYYY

## Thematic Rationale
[Why this batch, how it connects]

## Books
[Individual reports]

## Next Steps
[Gaps identified, what to acquire next]

Current Gaps (Priority)

URGENT - Missing Key Authors

Need More Coverage

Catalog Sources

Local Catalogs

• BPH Catalog (data/bph_catalog.csv) - 28,814 entries
• IA Catalog (data/ia_catalog.csv) - 9,000 entries

Online Sources

Session Tracking

Append session reports to curatorreports.md:

# Session [N]: [DATE] - [THEME]

## Acquired
| Title | Author | Pages | Book ID |
|-------|--------|-------|---------|
| ... | ... | ... | ... |

## Rejected
| Title | Reason |
|-------|--------|
| ... | Modern translation |

## Session Total: N books, N pages

Quality Flags

When issues are found:

• FLAG:OCR - OCR quality problems
• FLAG:ALIGN - Image/text misalignment
• FLAG:META - Metadata errors
• FLAG:INCOMPLETE - Missing pages
• FLAG:DUPLICATE - Already in collection