01
Register evidence
Accept explicitly supplied source content with a locator, stable source ID, trust flag, metadata and SHA-256 content digest.
ILAIOS Research & Data Factory
Research that keeps claims, sources and verification boundaries visible.
Research & Data Factory is a bounded implemented foundation in the ILAIOS repository. It records source provenance, separates proposed claims from verified facts, performs deterministic bounded numeric analysis and fails closed when evidence requirements are not met.
Provenance-first
A research output does not become a fact merely because a model produced it.
The current implementation does not autonomously crawl arbitrary external sources or imply general-purpose research coverage. It operates on explicitly supplied evidence and promotes claims only when configured trusted-source gates pass.
02
Propose a claim
Keep a claim separate from fact status and require it to reference known source IDs instead of relying on unsupported model narration.
03
Verify support
Require the configured minimum of trusted independent sources before a claim may become verified; the default bounded implementation requires two.
04
Fail closed
Unknown sources, duplicate evidence IDs, insufficient trusted support and invalid analysis inputs stop the workflow rather than silently weakening the gate.
05
Analyze deterministically
For bounded numeric inputs, retain a canonical values digest with count, minimum, maximum and mean so repeated analysis is reproducible.
06
Project verified knowledge
Only verified claims may project as Fact nodes, with Evidence nodes and explicit derived-from edges preserving provenance.
Connected knowledge