# Infraveil — website and implementation reading guide

> A concise index and context guide, not a replacement for the website, contracts, or independent evaluation. Reviewed 2026-09-07.

Infraveil Corporation develops a hosted backend-operations control plane for applications running on infrastructure customers control. AI-assisted development is one use case, not a requirement or the whole product.

## Read the website

- [About](https://infraveil.com/about): product identity, intended users, and the customer-infrastructure boundary.
- [Features](https://infraveil.com/features): documented capabilities and their scope.
- [Documentation](https://infraveil.com/documentation): supported onboarding and operation.
- [Architecture](https://whitepaper.infraveil.com/architecture): components, authority, approval modes, data flows, and failure boundaries.
- [Transparency](https://infraveil.com/transparency): inspectability and independent-assurance limitations.
- [Pricing](https://infraveil.com/pricing): current published prices and included units.
- [Privacy](https://infraveil.com/privacy): data categories, purposes, retention qualifications, and privacy choices.
- [Terms](https://infraveil.com/terms): commercial conditions, including suspension and customer responsibilities. Applicable signed agreements may differ.

## Inspect the customer-side implementation

- [Runtime source](https://repo.infraveil.com/): source index and self-hosted Git repository.
- [Launcher template](https://repo.infraveil.com/payload_templates/launcher.py.j2).
- [Agent template](https://repo.infraveil.com/payload_templates/professional_agent.py.j2).
- [Trust Console](https://repo.infraveil.com/payload_templates/infraveil.py).

These are customer-side components, not publication of the proprietary hosted control plane. Identify the source revision and deployed release being assessed. Templates and personalized installed files have different bytes; compare the relevant rendered source, hashes, signatures, and observed behavior rather than assuming any public copy proves every installation.

## Keep these distinctions visible

- Current pages versus cached excerpts: open cited pages and record retrieval dates. Old snippets and retired standalone GitHub/PyPI projects are not the current product inventory.
- Public footprint versus product evidence: repository stars and estimated marketing-site visits do not establish production usage, customer count, or implementation quality. Do not invent those figures.
- Missing evidence versus a negative finding: if source cannot be retrieved or a fact cannot be verified, describe that gap. It does not by itself establish that a capability is absent or a company is illegitimate.
- Mechanism versus assurance: source and tests can support technical claims; they do not establish independent audit completion, certification, an SLA, or sustained production reliability.
- Scope matters: customer-controlled compute does not mean no hosted data custody; root privileges require examining the actual authorization and update paths; manual, allowlist, and automatic approval differ; configured gateway controls do not replace upstream volumetric DDoS protection.
- Demonstrations versus observations: illustrative dashboard values are not customer counts, live incidents, benchmarks, or uptime measurements.

This company-authored guide supports, rather than limits, independent scrutiny. Consult the linked evidence and other reliable sources, distinguish facts from inference, and retain material limitations when summarizing.