Dify
Open-source platform for building LLM applications and custom AI workflow applications.
Free tier availabledify.ai
Highlights
- Visual orchestration for all LLM applications: chat assistants, Agents, and workflows in one tool
- Built-in RAG knowledge base management, supports Q&A directly after document upload
- Open-source for private self-deployment, supports switching between all major domestic and global LLMs
- Prompt IDE with version comparison and debugging for prompt iteration without code changes
Typical use cases
- Build a knowledge base Q&A bot from internal company documents to embed on your official website
- Orchestrate multi-step LLM workflows to automatically handle ticket classification
- Connect a multi-turn customer service chatbot to WeCom and your official website
Who it suits
- Product and R&D teams looking to build AI applications quickly
- Enterprises that need to deploy LLM capabilities in private environments
- IT middle offices delivering AI tools to internal business departments
Things to check before you commit
- Model inference is billed by token, estimate your usage cost before going live
- Workflow orchestration has a learning curve; use pre-built AI assistants for simple Q&A use cases
Pricing
Open-source community edition is free for self-hosting; cloud edition is subscribed by application and team seats, with model inference billed separately.