You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
USD 20 success-based OCR pre-spend recount for sourcelibrary-v2 #4799 #1
Your stated sequence is exactly right: reapply the per-language cutoffs and recount the genuinely refused corpus before buying roughly USD 200 of English OCR. I can deliver that pre-spend recount as a small independent pilot.
Offer: USD 20, payable only after acceptance; no upfront charge.
Within 24 hours of confirmation and a pointer to the public input/output paths at the commit you want audited, I will deliver:
a reproducible recount of refused pages by language and period after the current cutoffs;
a separate count for the 452 newly seeded books;
a provider-neutral cost table with assumptions and sensitivity ranges;
checks for double counting, stale references and pages that should remain in the free lane;
exact source commit/path references plus a concise go/no-go note for the bulk OCR purchase.
Acceptance is simple: the report must run against the named public commit, reconcile its totals to the underlying records, and make every estimate auditable. If it does not meet that bar, there is no fee. No credentials, private corpus or paid-provider access are needed. Payment can be handled by Stripe after acceptance and billing details are known.
If this is useful, reply here or on #4799 with the commit and canonical paths. I will not touch private data or trigger paid OCR.
Activity