Products 2026
Sinag
Making Philippine public documents searchable, from the scan up.
Sinag (“ray of light”) is a platform for public documents from Philippine government bodies, starting with Senate and House bills and extending to Supreme Court decisions and agency issuances. The goal is simple to state and hard to do: take documents that exist only as scans, and make them searchable without pretending the machine’s guess is the truth.
The problem with scans
Most Senate bills are photographs of paper. Some carry an OCR layer of unknown quality; some carry none. Either way, the text is a guess until a person checks it. Sinag keeps the original scan archived and linked next to the transcribed text, so a citation can always be verified against the source.
Sources differ wildly. Some have APIs or semi-structured data; some are bot-protected sites that need careful crawling; some are case-law databases; some are plain listings. Each source gets its own ingestion path, and the pipeline normalises them into one document model.
From scan to corpus
The pipeline crawls, archives the original, runs OCR, and routes the result through proofreading. Corrections are stored as data, not as overwritten text, so a later OCR model can be compared against human edits. The resulting corpus serves two audiences: a search UI for researchers, and semantic search / RAG endpoints for AI tooling.
Why it matters
Legislative and judicial documents are public but not usable. Full-text search across pending and historical documents is not offered by any free or paid service in one place. Sinag is an attempt to build that index properly — with the provenance that legal research requires.