Search the corpus
Query 1.27M federal records and 2.5M Alberta provincial records by topic, recipient, or program. Hybrid retrieval — keyword and semantic — every result is auditable to its source row.
The question is concrete: where does Canada's public money go, and to whom?
Glassbox traces federal grants, Alberta contracts, charity flows, and sole-source amendments across millions of records. Every figure cites its source row. Calibrated language. Audit-ready.
We surface twelve named patterns of public-spending behaviour — circular charity flows, contract growth far beyond the original bid, recipients that vanished after the money flowed, departments dominated by a single supplier. Each pattern is a search you can run against the corpus, with severity bands, evidence arrays, and calibrated next-step recommendations.
Built on Alberta's TRACE methodology, extended to the federal corpus, packaged as a public interface. Open source, MIT, with audit tokens chained to source data tokens — so the “why this match?” trail can always be walked back.
Pick a pattern. Glassbox surfaces every record matching it, in calibrated language, with full citation.
Three primary capabilities. One coherent product.
Query 1.27M federal records and 2.5M Alberta provincial records by topic, recipient, or program. Hybrid retrieval — keyword and semantic — every result is auditable to its source row.
Paste a solicitation draft to surface duplication risk, recipient concentration, and calibrated-language issues before publishing — not after audit.
Every result, every score, every recommendation carries an audit token. Downloadable JSON. Independently verifiable.
Every draft is evaluated against four dimensions. The composite score is the recommendation.
The closest existing record has similarity 0.83 with the draft.
Top three recipients account for 71% of comparable funding.
Two flagged phrases against the calibrated regex set.
Audit tokens wrap every score with full provenance. Four governance tiers — input filtering, contextual modeling, output gating, audit sealing — each with passing gates, evidence count, and retrieval timestamps. Independently verifiable. Downloadable as JSON.
View the methodologyGlassbox runs every output against a published regex set, a citation discipline, and an audit-token schema. The validator is open. The scoring is reproducible. The audit trail is the product.
View the full methodology — including the calibrated lexicon, the data landmines, the AIA structural correspondence, and a live try-it-yourself validator.