Leaderboard · run 2026-11-02
What AI coding agents install, in 20 devtool categories
We gave Claude, GPT and Gemini real build tasks that never name a product, and counted what each one actually installed. 600 answers in all. Click a category to see the prompts and the answers.
| Category | Most installed | Share | Claude picks | GPT picks | Gemini picks | Runs |
|---|---|---|---|---|---|---|
| Auth | Auth.js / NextAuth | 27% | Auth.js / NextAuth 40% | Auth.js / NextAuth 20% | Clerk 30% | 30 |
| Postgres hosting | Neon | 67% | Neon 80% | Neon 60% | Neon 60% | 30 |
| ORM / query builder | Drizzle | 53% | Drizzle 60% | Drizzle 60% | Drizzle 40% | 30 |
| Payments | Stripe | 80% | Stripe 80% | Stripe 80% | Stripe 80% | 30 |
| Transactional email | Resend | 53% | Resend 40% | Resend 60% | Resend 60% | 30 |
| Hosting / deploy | Vercel | 40% | Vercel 40% | Vercel 40% | Vercel 40% | 30 |
| Object storage | AWS S3 | 57% | AWS S3 60% | AWS S3 60% | AWS S3 50% | 30 |
| Background jobs / queues | BullMQ / Bull | 40% | BullMQ / Bull 40% | BullMQ / Bull 40% | BullMQ / Bull 40% | 30 |
| Workflows / cron | Cron library | 40% | Cron library 40% | Cron library 40% | Cron library 40% | 30 |
| Feature flags | PostHog flags | 33% | Vercel Flags 30% | LaunchDarkly 30% | PostHog flags 50% | 30 |
| Error monitoring | Sentry | 100% | Sentry 100% | Sentry 100% | Sentry 100% | 30 |
| Product analytics | PostHog | 93% | PostHog 80% | PostHog 100% | PostHog 100% | 30 |
| Logging / observability | Grafana / Loki / Prometheus | 40% | OpenTelemetry 50% | Grafana / Loki / Prometheus 60% | OpenTelemetry 40% | 30 |
| Search | Meilisearch | 37% | Meilisearch 50% | Meilisearch 30% | Meilisearch 30% | 30 |
| Vector database | pgvector | 33% | pgvector 40% | pgvector 60% | Pinecone 20% | 30 |
| LLM observability | Langfuse | 33% | Langfuse 40% | Langfuse 40% | LangSmith 50% | 30 |
| Headless CMS | Sanity | 37% | Sanity 20% | Sanity 60% | Git-based CMS 30% | 30 |
| Realtime / websockets | Socket.IO | 33% | Ably 20% | Socket.IO 40% | Socket.IO 40% | 30 |
| E2E testing | Playwright | 100% | Playwright 100% | Playwright 100% | Playwright 100% | 30 |
| UI component library | shadcn/ui | 63% | shadcn/ui 70% | shadcn/ui 60% | shadcn/ui 60% | 30 |
Findings
- Supabase is mentioned, never installed: the Postgres host AI agents pick
- Drizzle vs Prisma: which ORM Claude, GPT and Gemini install
- Sentry and Playwright: the only tools AI agents install every time
- Which email API do AI agents install? Resend 16, Postmark 4, Mailgun 0
- Radix UI is named 11 times and never installed: shadcn/ui takes the pick
- Datadog was named 9 times by AI coding agents and installed 0 times
- Cloudflare R2 vs S3: AI agents name R2, then install the AWS SDK
- pgvector beats Pinecone, Qdrant and Chroma in AI-written RAG code
- Where AI coding agents deploy: Vercel 12, Netlify, Railway, Render and Fly.io 0
Share = the product the answer installed first, out of all answers in that category. Samples are small (about 10 answers per model per category), so read the bands on each category page. How we measure.
Missing a category or a tool?
A tool we missed, a pick that looks wrong, a prompt we should add? Tell us. We read every note and fix the data in the next run.
Is your product losing these picks?
Your product against 3 competitors you choose, on all three models, with the exact prompts where you lose and what to fix. $19, delivered within 24 hours.