kekzl/imp
From-scratch C++23/CUDA inference engine for the NVIDIA RTX 5090 (sm_120a). The best single-GPU backend for agentic AI: tool calling, long-context loops, reasoning and concurrent sub-agents. Decode beats llama.cpp b9976 by 42-48% on dense GGUF (measured 2026-07-12), at-or-ahead of vLLM on NVFP4. 100% written by Claude Code.
36
Stars
2
Forks
31.6
Radar Score
Stable
Momentum
Scoring Breakdown
Impact15.2
Velocity31.4
Health58.3
Growth Metrics
7-Day Stars
+0
30-Day Stars
+0
Acceleration
x
Language
Cuda
Popular Alternatives & Competitors
Top projects in the same category (AI Coding Tools) — compare side by side.
🛡️ Enterprise Readiness & Commercial Safety
Organic Growth
95%
High Confidence
Growth pattern looks organic.
License
Permissive (Commercial Friendly)
Safe for commercial use, modification, and distribution.
Health Grade
A
Based on maintenance, issue response, and release cadence.
Radar Score
31.6
Composite of Impact, Velocity, and Health.
Disclaimer: This is an automated heuristic assessment based on public GitHub signals. Always perform your own security audit, dependency review, and legal compliance check before production deployment.
Share & Embed
Display live Radar metrics in your documentation, website, or social profiles.