Verification & Provenance
Methodology & Claim Boundaries
Last updated: 2026-10-02
Launchfiles is the launch operator for AI-built products: it takes a product to its first paying customers and keeps them, through agents acting inside a policy you approve, with a receipt for every public act. Every Launchfiles radar separates deterministic code measurement (HTTP status codes, rendered DOM geometry, JSON-LD graphs, and Search Console impression/CTR deltas) from model-assisted drafting. A language model never invents scores, rankings, or traffic forecasts.
1. Three-Pillar Authority & Trust Scoring Architecture
On-Site Conversion Trust
- Transparent /pricing (HTTP 200) with plain-English cancellation terms
- 14-day refund window & 3-step billing timeline beside checkout CTA
- Verifiable publisher identity (/about, support@) & dated /security subprocessors
- Interactive above-the-fold product proof + zero fake vanity logos
Off-Site Cold-Start Authority
- Public GitHub / package registry or vertical registry presence
- Live directory listings verified for HTTP 200 & real dofollow vs /go/ redirects
- Claimed G2, AlternativeTo, and Trustpilot entity profiles
- 4-layer canonical naming consistency across domain, schema, and profiles
AI & LLM Citation Readiness
- AI search bots (OAI-SearchBot, PerplexityBot, ClaudeBot) allowed + SSR HTML
- Honest /vs/[competitor] comparison tables with explicit trade-off concessions
- 40–60 word BLUF answer capsules in the top 30% of key pages
- Valid Organization + SoftwareApplication sameAs JSON-LD & /llms.txt
2. Deterministic Signal vs. Model-Assisted Output Matrix
| Diagnostic Surface | Deterministic Code Measurement | Human-Reviewed Work Order Output |
|---|---|---|
| On-Site Trust (Signals 1–8) | Live HTTP 200 probes on /pricing, /security, /methodology, /changelog, and DOM refund/timeline proximity checks. | Drop-in StepsTrialAndGuaranteeBlock and self-serve trust route PR bundles (WO_1–WO_6). |
| Off-Site Authority (Signals 9–15) | Outbound link parser inspecting rel="nofollow|ugc|sponsored" and /go/ redirect wrappers on live directory listings. | Verified directory submission kit and 3-question customer review outreach templates (WO_7). |
| AI / LLM Citation Readiness (Signals 16–20) | robots.txt AI bot parser, SSR payload check, JSON-LD sameAs validator, and /llms.txt Spec v2 check. | Honest /vs/[competitor] comparison tables and entity-graph.jsonld patches (WO_8–WO_9). |
3. Explicit Limitations & What Launchfiles Cannot Do
- No guaranteed rankings, traffic, or revenue: Passing all 20 deterministic trust and citation signals removes technical and conversion friction, but search rankings, LLM citations, and buyer conversion depend on market demand and competitor activity.
- Screenshot-bound UX previews: Free Design UX Radar runs inspect the rendered first screen and DOM of a single public URL; they do not observe private authenticated funnels or live session heatmaps unless connected inside an invited workspace.
- Human approval boundary: Launchfiles never auto-submits to external directories, never fabricates user reviews, and never merges code into your repository without your explicit review and sign-off.
External Standards & Benchmark Citations
Diagnostic observations and crawler discovery checks align with published open standards: W3C WCAG 2.2 for visual contrast and target geometry, IETF RFC 9309 for Robots Exclusion Protocol compliance, Schema.org SoftwareApplication for structured entity graphs, and Google Search Central Title Link Guidelines for search snippet evaluation.
Questions about our scoring formulas or verification gates? Contact socials@launchfiles.app or review our Security posture and Pricing boundaries.
Launchfiles is a trade name of Launchfiles Operations, auto-entreprise (French sole proprietor).
Launchfiles Digital Services
SIREN: Registered Entity
APE: 73.11Z · Activity start: 08/09/2025
SIRET: Registered Entity
VAT: TVA non applicable, art. 293 B du CGI (franchise en base de TVA, no VAT number)
Operator: Launchfiles.