Turbocharge Claude Code, Cursor, Codex, Gemini & every coding agent: faster, cheaper, with contextual understanding specific to your codebase.
TypeScript ai-agentsanthropicclaude-codeclicode-graphcodexcontext-engineeringcursordeveloper-toolsgeminiknowledge-graphllmmcpmcp-serveropen-sourceopenaitree-sitter
docs: restore hero mean table (peak line + ~29% cost / ~21% wall averages)
Shrish Dwivedib254cf0a9cdc
diff --git a/README.md b/README.mdindex 043d932f..ee2f10a0 100644--- a/README.md+++ b/README.md@@ -17,7 +17,13 @@### Up to **4× cheaper** and **3× faster**, with better or no loss of correctness.-<sub>Measured on real tasks (real questions + real merged PRs) in popular open-source repos: PocketBase, ollama, Excalidraw. Same agent, same file tools; only the context differs. [Full per-repo numbers ↓](#tested-on-your-popular-repos)</sub>+| Averaged across real tasks | Standard Claude Code | With graft |+|---|---|---|+| Cost | baseline | **~29% less** |+| Wall-clock | baseline | **~21% less** |+| Correctness | baseline | **matched** (reproduced the merged PRs) |++<sub>Measured on real tasks (real questions + real merged PRs) in popular open-source repos: PocketBase, ollama, Excalidraw. Same agent, same file tools; only the context differs. The "up to 4×" figures are the biggest single-task wins; the table is the average. [Full per-repo numbers ↓](#tested-on-your-popular-repos)</sub></div>