Reproducibility is the load-bearing check on every claim Prodia makes. These kits let qualified researchers rerun the core experiments end-to-end, against the same frozen corpora used under Foundation v1.0. Access is administrator-approved to protect the integrity of the record.
Approved researcher, partner or evaluator? Open the Research Console →
Prodia treats every sealed result as a claim under test. The kits below provide controlled protocol materials, expected-result fixtures and signed evidence summaries sufficient for an independent researcher to evaluate selected published claims.
Frozen corpora, signed artifacts, no post-hoc edits.
Selected protocol materials and expected-result fixtures accompany every kit.
Kits publish what would count as a failed reproduction.
Reproduces the predictive relationship between dependency structure, verification reliability and observed capability compounding.
Reproduces the structural-compounding forward prediction on the supplied blinded corpus within the published tolerance.
Reproduces autonomous generation and abstraction reuse under independent verification, from initial inputs to sealed artifacts.
Reproduces the compounding chain and rollback integrity numbers reported under Foundation v1.0.
Implementation-level detail is reserved for the Research Console under an approved research-access agreement.
Reproducing a result from a kit is meaningful. It is not the same as an independent external replication. Prodia is explicit about which level any given piece of evidence has reached.
Prodia's own runs against the frozen corpus, under the sealed scientific loop. Highest control, lowest independence.
An external team executes the supplied kits against the same frozen corpus. Tests kit clarity and reproducibility of the published result.
An external group produces its own implementation, corpus and verification process. External implementation, external corpus, external verification — the strongest form of evidence and the ongoing goal.
Reproducing a claim from an approved kit is meaningful evidence; it is not equivalent to independent external reimplementation.
Tell us who you are and what you intend to test. We prioritise requests that describe a concrete replication or falsification plan. You'll receive a decision by email.
Prefer to browse the record first? See the evidence registry or the open questions.
