Skybrary scale
Corpus ops
Software v1.35.0 ships corpus doctor, status, printable board, kit, and batch manifests. Builds a legal subset only - not every book, not free commercial broadband, never pirate mirrors.
Legal fail-closed
Every import requires an open license string. Unknown or commercial-restricted licenses are refused. Prefer primary hosts (Gutenberg, open OA) and record provenance for partners and regulators.
Operator commands
# Product surface (top-level alias) python -m skycache corpus doctor python -m skycache corpus status python -m skycache corpus export --out data/ops/corpus-board.html python -m skycache corpus kit --out data/corpus-kit # Offline demo batch (samples + test fixtures only) python -m skycache corpus sample-manifest --out data/corpus-batch-demo.json python -m skycache corpus batch \ --manifest data/corpus-batch-demo.json \ --allow-local --ingest --data-dir data # skybrary corpus ... still works; one-shot tools python -m skycache skybrary import-folder ./open-texts --license "public domain" --ingest python -m skycache skybrary provenance --data-dir data
Batch job types
- folder - local TXT/MD/HTML/EPUB with required license
- gutenberg_catalog - rate-limited catalog snapshot (fixture or operator file)
- oa_science - open-access science catalog (license-gated)
- open_url - single allowlisted open HTTPS URL
Full runbook: skybrary-corpus-import.md.