Scoring methodology
Rubric version 1.0. Weights and freshness rules below are public; per-tool component scores are a Pro feature.
Weighted overall (default ranking)
| Factor | Weight | What it measures |
|---|---|---|
| voice_coder_fit | 25% | How well the tool supports voice coding, speech-to-code, hands-free, or realtime voice agent workflows. |
| builder_fit | 25% | How well it serves builders shipping agents/apps (APIs, DX, docs, extensibility). |
| agent_maturity | 20% | Production readiness for agentic use: tool use, multi-step, reliability, ecosystem. |
| freshness_score | 15% | Derived from last_release_at using freshness_rules. |
| value_score | 15% | Pricing clarity and accessible entry (see value_score_guidance). |
overall = Σ (factor × weight), rounded to 2 decimals. Overall is free; full component breakdowns on tool pages are a Pro feature.
Freshness rules
Derived from last_release_at relative to publish date:
| Days since release | Score |
|---|---|
| ≤ 14 days | 10 |
| ≤ 30 days | 8 |
| ≤ 90 days | 5 |
| ≤ 180 days | 3 |
| older | 1 |
Value / pricing clarity
Rate 0-10 on pricing clarity + accessible entry point: clear public pricing + free/freemium tier near 8-10; opaque enterprise-only nearer 3-5; unknown/unclear 2-4.
Ties & category rankings
Sort by overall descending within each category. Ties broken by tie_breakers in order (higher score wins; name ascending as final).
Tie-break order: freshness_score, agent_maturity, voice_coder_fit, name.
Human judgment
Fit scores (voice, builder, maturity, value) are editorial estimates based on public product docs and ecosystem reputation. Freshness is algorithmic. Re-check dates before making buying decisions.
Sponsorship & affiliates
Paid Featured slots never alter organic order. Outbound “Visit site” links go through go/<id>.html (affiliate URL when active, else website) with UTM params.