dify vs ART
A neutral, constraint-first comparison - live GitHub stats and typed relationships, not marketing.
| dify | ART | |
|---|---|---|
| Tagline | Production-ready platform for agentic workflow development | Agent Reinforcement Trainer |
| Stars | 148k | 10k |
| Forks | 23k | 937 |
| Open issues | 851 | 121 |
| Language | TypeScript | Python |
| License | Other | Apache-2.0 |
| Last pushed | Jul 7, 2026 | Jul 7, 2026 |
| Categories | AI Agents | AI Agents, Model Training |
dify
Dify is a production-ready platform designed to facilitate the creation and management of agentic workflows. It supports TypeScript, Python, and leverages frameworks like Next.js.
TypeScript
ART
OpenPipe's ART is an open-source RL framework for training multi-step agents using GRPO, designed to enhance the reliability of LLMs through experience learning.
Python