vercel/ai

What happens to a stranger's pull request here, in the Porchlamp snapshot of 14 October 2026.

Outside pull requests to ai got a first human reply in a median of 1.5 hours (158 pull requests, 10 April to 7 October 2026). The share with no reply within 14 days has too small a sample to measure (0 pull requests counted; 10 needed). It ranks 312 of 3,833 overall. It is not ranked for first-time contributors: not enough evidence to show under this lens.

Rating

  • Score: 64.93
  • Coverage: 78% · Ceiling B
  • Tier: B (capped)
  • Quadrant: Sweet Spot
  • Ecosystem: npm

Capped: this tier can go no higher until more of the project's inputs are measured. It reflects how much data there is, not how good the score is.

Five pillars

Score out of 100, and how much of the pillar was measured
PillarScoreCoverage
Responsiveness96.140%
Criticality66.0100%
Governance68.0100%
Rigor50.9100%
Accessibility21.050%

Rank in each lens

LensStanding
BalancedRank 312 of 3,833
First-timerNot ranked: Not enough evidence to show under this lens.
Career capitalRank 301 of 823
Senior systemsRank 174 of 1,206
AI-assistedRank 498 of 3,929
The measured parameters
ParameterValueCountWindowWhat it measures
Time to First Response1.495415810 April to 7 October 2026Median hours from an outside pull request opening to its first substantive human reply.
Decision Rate (90d)Not shown: Insufficient sample size——Share of outside pull requests that were merged or closed within 90 days.
First-timer Merge RateNot shown: Insufficient sample size——Share of first-time outside authors' decided pull requests that were merged.
PR Ghost Rate (14d)Not shown: Insufficient sample size——Share of outside pull requests with no maintainer response, merge or close within 14 days.
Time to Merge (p90)47.700316310 April to 7 October 2026Hours within which 90% of merged outside pull requests were merged.
Dependents185——Packages in the registry that depend on this project, counted once each.
OpenSSF Criticality0.5557——OpenSSF criticality score from 0 to 1, computed from the project's public activity signals.
Monthly Downloads110341313——Downloads from the package registry in a month.
Legal IP RegimeNONE——The contributor licence regime: DCO, none, a foundation CLA or a corporate CLA.
Elephant Factor3484—The fewest organisations that account for half of the last year's commits.
Promotion LadderAD_HOC——How openly a contributor can become a maintainer, from closed to verified.
Pipeline Defenses50.6173243—How well the project's build and review pipeline defends against bad changes, 0 to 100.
AI Policy PostureREGULATED——Whether the project has a written policy on AI-assisted contributions.
Substantive Review Rate0.058625610 April to 7 October 2026Share of decided pull requests that received a substantive review.
Newcomer Issue Supply0——Open, unassigned, up-to-date issues marked for newcomers.
Newcomer ConversionNot shown: Insufficient sample size——Share of first-time authors whose first pull request merged within 180 days.
Newcomer Return RateNot shown: Insufficient sample size——Share of those merged first-time authors who sent a second pull request within 90 days.
Developer Experience7010—Developer experience, 0 to 100: half CI run time, half setup signals.

Similar projects

The ranked projects nearest this one in the same ecosystem.

Evidence

  • Snapshot 2026-10-14, data as of 13 October 2026.
  • Record hash: 3bbe3bb8d4f3978d8652abbfc205b2afc331d9de2fed25f48b7c2f0554ef5a71
  • The full record is in data/2026-10-14/repos/80.json, under the key vercel/ai.
  • How every number is measured: Porchlamp methodology.

Is something wrong?

Open an issue about this record and say what looks wrong.