FrameworksPreviewThis index is in active development: we are still adding agents, refining attribution signatures, and verifying data against sources. The date in the header is when key figures were last re-checked. Numbers can move as coverage improves.
Where each agent's work concentrates across the major frameworks.
19%
Next.js draws the most agent work
Share of classified sampled volume across 21 frameworks
68%
Devin leads Spring
The strongest single-agent concentration
6.8%
Of sampled PRs matched a framework
The rest of the sample has no framework signal
Framework Tilt
Window: 2026-08-03 to 2026-08-23.
| Next.js | 21% | 20% | 8% | 27% | 10% | 15% | 19% |
| React | 32% | 20% | 13% | 22% | 7% | 5% | 16% |
| Flutter | 14% | 16% | 12% | 31% | 5% | 22% | 14% |
| .NET | 9% | 13% | 59% | 4% | 13% | 2% | 6% |
| FastAPI | 25% | 8% | 12% | 8% | 2% | 44% | 6% |
| Electron | 11% | 12% | 24% | 2% | 51% | 0% | 6% |
| Astro | 33% | 18% | 21% | 14% | 4% | 11% | 5% |
| Vue | 17% | 12% | 20% | 9% | 9% | 33% | 4% |
| Spring | 4% | 9% | 13% | 5% | 68% | 0% | 4% |
| Express/Nest | 14% | 8% | 17% | 22% | 5% | 34% | 3% |
| Angular | 6% | 5% | 13% | 8% | 67% | 1% | 3% |
| Svelte | 15% | 22% | 33% | 10% | 3% | 17% | 3% |
| SwiftUI | 32% | 32% | 8% | 13% | 14% | 1% | 2% |
| React Native | 36% | 17% | 6% | 23% | 9% | 8% | 2% |
| Flask | 13% | 6% | 29% | 21% | 8% | 22% | 2% |
| Django | 12% | 3% | 18% | 55% | 12% | 0% | 1% |
| Tauri | 21% | 38% | 7% | 25% | 10% | 0% | 1% |
| Rails | 29% | 16% | 12% | 12% | 31% | 0% | 1% |
| SvelteKit | 17% | 8% | 48% | 17% | 0% | 6% | 1% |
| Streamlit | 10% | 26% | 33% | 5% | 23% | 3% | 1% |
| Laravel | 23% | 35% | 23% | 6% | 3% | 10% | 1% |
Cells are each agent's share of the framework's sampled agent PRs; Volume is the framework's share of all classified samples. Click a header to sort; hover a cell for the raw count. Equal-size samples per agent, so this is portfolio tilt, not market share.
Movers
2026-08-24 vs 2026-08-17Tilt Over Time
Window: 2026-06-11 to 2026-08-24.
Framework shares are portfolio tilt, not market share: every agent contributes an equal-size sample, so a small agent can lead a framework it focuses on. For market-share-style adoption with a true baseline, see Languages. Method details on the methodology page.