Ask VitContactAsk VitGet Started
ENIT

Open source · Domain-agnostic · Trace-first

Orchestrate AI.
Preserve the evidence.

Motus is the trace-first evidence foundation for intelligent systems — from open execution to governed enterprise access and long-term verification.

Hold the rule and you can check any point without being shown the rest.

Leonardo da Vinci, Paris Manuscript B, fol. 83v · Bibliothèque de l'Institut de France, c. 1487-90 · plate

Today most AI interactions have two characters.

User asksAI answers

Motus introduces the auditor.

When AI decisions matter, a third party may need to inspect what happened between the question and the answer — and verify the evidence afterwards.

Draft — pending legal review

High-risk under the EU AI Act (Annex III): automated creditworthiness evaluation of a natural person.

Can John Smith’s €35,000 loan application be approved automatically?

No. The verified debt-to-income ratio exceeds Acme Bank Europe’s 43% automatic-approval threshold, so a credit officer — not the system — makes the final call.

Three products. One evidence model.

Open source

Motus Continuum

Understand what happened.

A trace-first execution runtime and inspection layer for AI workflows.

  • Graph execution
  • Causal evidence
  • Open verification

Free. Open. Embeddable.

Enterprise

Motus Perpetuum

Keep the evidence trustworthy over time.

Governance, retention, assurance and audit workflows for organisations running consequential AI.

  • Evidence governance
  • Retention & recovery
  • Audit & assurance

Enterprise governance and assurance.

Enterprise gatewayIn development

Motus Limen

Control enterprise access to AI.

One OpenAI-compatible endpoint that applies company policy, routes requests to approved models and records verifiable evidence with Motus.

  • Access control
  • Approved model routing
  • Verifiable evidence

One endpoint. Policy enforced. Usage recorded.

Install it. Run it.

Motus embeds into existing Python systems without imposing an LLM, agent framework or domain model.

terminal
$ pip install vitruvyan-motus

  Successfully installed vitruvyan-motus-0.9.0

$ python run.py

  node retrieve → 5 citations
  node classify → intent=support
  node respond  → effect committed

  ✓ run complete — trace written to run.json
PythonApache 2.0No LLM dependencyDomain agnostic

AI produces decisions.
Motus preserves the evidence behind them.