Inputs
Primary source collection
Official site, docs, pricing, privacy, security, repository activity, release notes, and migration documentation.
Methodology
How ASR verifies, scores, and explains privacy-first software choices
The recommendation engine should feel auditable. This page shows what counts as evidence, how weighting works, and where the editorial layer gets intentionally strict.
05scoring dimensions
03review stages
0sponsor-driven verdicts
Scoring
Five weighted dimensions
Privacy posture, openness, maintenance, usability, and business readiness. No single idealist metric decides the outcome.
Guardrails
Editorial rules that keep the system honest
Never call something best without a use case, never hide migration pain, and never let sponsorship rewrite ranking logic.
Scoring dimensions
What the system actually measures
These are the dimensions users should understand before trusting any recommendation coming from the app.
Privacy posture
Telemetry defaults, encryption claims, hosting control, data exposure, and what the vendor does not say clearly enough.
Openness
License reality, source visibility, self-hosting paths, API quality, and whether openness actually survives contact with deployment.
Maintenance
Release freshness, documentation quality, issue hygiene, roadmap credibility, and whether the product feels alive.
Usability
Onboarding friction, cross-platform quality, workflow fit, migration complexity, and the real cost of re-training a team.
Business readiness
Team support, admin controls, billing clarity, import-export tooling, compliance cues, and buyer confidence.
Review pipeline
From raw evidence to a verdict readers can actually use
Stage 01
Collect evidence from primary product surfaces.
ASR starts with official docs, pricing, privacy language, changelog reality, and migration constraints before any score exists.
- Record hard product facts first.
- Separate vendor claims from verified constraints.
- Build the scoring model on raw evidence, not recycled summaries.
Why this matters
The app will only be trusted if the main domain proves the logic first.
That is why methodology is not a legal footnote here. It is product design, editorial trust, and future monetization discipline all at once.