YC batches · batch ranking
YC W23 — Winter 2023YC W23
The 8 Winter 2023 companies we track, ranked by evidence-graded agent readiness — can an agent actually drive this product? — then built-in AI and PA Score.
This is the YC W23 companies we track, judged on evidence — not a batch-wide census (yet). Coverage grows batch by batch. Scores come from each product’s arena verdicts — see methodology.
| # | Product | Agent-ready? | Built-in AI | PA Score |
|---|---|---|---|---|
| 1 |  LiteLLMModel Gateways & RoutersOpen source | 61/100 | unscored | 34/100 |
| 2 |  ExtendDocument Extraction APIs | 58/100 | 34/100 | 33/100 |
| 3 |  FernDeveloper Docs Platforms | 56/100 | 47/100 | 43/100 |
| 4 | Trigger.devDurable Execution EnginesOpen source | 55/100 | 26/100 | 39/100 |
| 5 |  LangfuseLLM Evals & ObservabilityOpen source | 52/100 | 19/100 | 34/100 |
| 6 | PylonAI Customer Support Agents | 51/100 | 51/100 | 31/100 |
| 7 |  NumeralSales Tax Automation | 43/100 | 21/100 | 20/100 |
| 8 |  HeliconeLLM Evals & ObservabilityOpen source | 43/100 | 24/100 | 34/100 |
A product tracked in more than one arena appears once, in the arena where its PA Score is highest. Unscored means no applicable cells — never zero. Batch stamps verified against YC’s public directory by website domain, never by name.