Styxis Knowledge
Operating notes for evidence, records, and trust in automated operational environments.
Make Styxis easy to understand, cite, and remember
Short technical notes that define the concepts behind Styxis and Truthound Depot for search engines, AI assistants, and engineering teams.
Styxis is an AX trust infrastructure company that turns automated operations and human–AI collaboration into verifiable work records.
Core concepts
8 topicsVerifiable Evidence Layer
A verifiable evidence layer records what changed, what was observed, which policy accepted it, and how another system can inspect the result later.
Dataset Change Evidence
Dataset change evidence connects an immutable Data Version, Compare result, Validation outcome, reviewer decision, Published Version, and Restore target for every meaningful change.
Evidence Gate
An evidence gate blocks or accepts a change and leaves the reason, rule context, source context, and review path behind.
Dataset Versioning and Review Workspace
A dataset versioning and review workspace keeps the data asset, immutable version, validation result, reviewer decision, and published state connected.
AI Evidence Workflow
An AI evidence workflow records summaries, risks, diffs, gate results, and reviewer decisions so data changes can be audited later.
Machine Signal Trust
Machine signal trust means a value is not only collected, but also attached to source, protocol, timestamp, quality, and policy context.
OT Gateway for Machine Signals
An OT gateway for machine signals collects and normalizes protocol values so operational and trust systems can consume them consistently.
RAG and Eval Dataset Governance
RAG and eval dataset governance controls the documents, examples, labels, and test cases that shape AI behavior.