← Back to all frameworks NLP

Dify

Open-source LLM application platform — visual agent + RAG workflows

What it is

Dify is an open-source platform for building production LLM apps. Drag-and-drop workflows, RAG over your docs, agent definitions, prompt management, observability — all behind a clean self-hostable UI.

How Vaaani uses it

  • Letting non-engineers on a customer team build their own AI workflows
  • Quickly stitching RAG + tool-calling without writing orchestration code
  • Self-hosted alternative when the customer can't ship data to a SaaS
  • Multi-tenant LLM ops with per-app rate limits and billing

Why it makes the cut

Sometimes the customer wants to maintain the AI worker after I deliver. Dify gives them a UI to do that — keeping me out of every small prompt change.

Sample code

# docker-compose.yml — self-host Dify
services:
  api:
    image: langgenius/dify-api:latest
    environment:
      DB_USERNAME: dify
      VECTOR_STORE: qdrant
  web:
    image: langgenius/dify-web:latest
    ports: ["3000:3000"]

Related in the Vaaani stack

Have a project that needs Dify?

30-min discovery call. You describe the busywork; I map it to an AI worker and a budget.