# Veriom Documentation

> Veriom derives an evidence-backed architectural model from code, infrastructure, delivery configuration, and submitted technical evidence. Every canonical relationship records why Veriom believes it exists.

## Start here

- [Full documentation](https://docs.veriom.ai/llms-full.txt): Complete public product and API documentation in one file for full-context retrieval.
- [How Veriom works](https://docs.veriom.ai/docs): The product model from technical signals to architectural truth and review-ready decisions.
- [Quickstart](https://docs.veriom.ai/docs/quickstart): Connect evidence, run an audit, inspect provenance, and export the result.
- [Architectural truth](https://docs.veriom.ai/docs/architecture/architectural-truth): Canonical entities, evidence-backed relationships, confidence, and contradictions.
- [Context engine](https://docs.veriom.ai/docs/architecture/context-engine): How bounded context packs support specialist agents without replacing canonical truth.
- [Agent system](https://docs.veriom.ai/docs/workflows/agent-system): Typed collaboration between security and architecture roles.
- [API overview](https://docs.veriom.ai/docs/api): Versioning, authentication, idempotency, typed errors, and asynchronous workflows.
- [API reference](https://docs.veriom.ai/docs/api-reference): Endpoints generated from Veriom's canonical OpenAPI document.

## Core definition

Architectural truth is Veriom's evidence-backed, time-aware model of what a system contains, how its parts relate, and how confidently each relationship can be proven. Diagrams are views of that model, not the source of truth.