Dify

What it does

Dify takes in prompts, model configurations, tools, data sources such as files, websites, online documents, and drives, and plugin integrations, and produces published AI applications. Builders assemble visual workflows and agents, prepare knowledge bases through repeatable pipelines that extract, clean, chunk, and index content, and test retrieval before connecting it to agents or workflows.

Teams use Dify to design a visible execution path for how an app retrieves data, makes decisions, calls tools, requests human input, and finishes tasks, then publish the result as a hosted web app, API endpoint, embed, or MCP-compatible tool. Logs, feedback, annotations, and usage data support ongoing improvement. Reusable plugins from the Marketplace let teams share approved model providers and tools across projects.

Dify suits developers and business teams building production AI applications, including those needing a no-code path. Dify Cloud offers a free Sandbox plan and paid Professional and Team workspaces billed per workspace per year, with message credits and quotas. Self-hosting the Community Edition uses Docker and an Apache-2.0-derivative license, while Enterprise adds SSO, RBAC, audit logs, and custom pricing. Model API costs and quotas still apply.

What it can do

• Build agentic workflows with visual building blocks and AI prompts that control retrieval, decisions, tool use, and human input
• Build agents by chatting or configuring them manually, then use them as apps, embedded widgets, or workflow nodes
• Create knowledge pipelines that extract, clean, chunk, index, and test retrieval over files, websites, online documents, and drives
• Install model providers, tools, data sources, and MCP integrations from the Marketplace or define custom plugins
• Publish apps as hosted web experiences, API endpoints, embeds, or MCP-compatible tools
• Monitor apps with logs, feedback, annotations, latency, and usage data

Works with

• MCP (Model Context Protocol) integrations
• Marketplace model providers and tools

Example workflows

Customer support knowledge assistant

Illustrative example. Inputs: company help documents and product pages. Steps: create a knowledge pipeline to extract, clean, chunk, and index the files; test retrieval; build an agent that answers from the knowledge base and escalates to a human when needed; publish as a web app. Output: a hosted assistant that cites retrieved content and logs conversations for review.

Internal approval workflow with tool calls

Illustrative example. Inputs: a request form submitted through an API endpoint. Steps: build a visual workflow that classifies the request, calls an approved tool from the Marketplace, pauses for human input, and completes the task; publish as an API endpoint. Output: a routed decision and an audit trail in logs.

Before you choose

• Dify Cloud quotas on apps, knowledge documents, storage, and message credits vary by plan; check the pricing page before scaling.
• Self-hosted Community Edition is source-available under an Apache-2.0-derivative license; Enterprise needs a commercial license and custom pricing.
• Model provider costs and API rate limits apply separately to Dify subscriptions.

Listed under

• Coding
• App builders
• AI agents
• Developers
• Enterprise
• Enterprise AI
• LLM app development
• LLM app platform

Get started

Free Sandbox plan available; self-hosting via Docker is also documented.

1. Create a Dify Cloud account at cloud.dify.ai, or install the Community Edition with Docker Compose following the documentation.
2. Connect a model provider from the Marketplace or add your own API key once Sandbox message credits run out.
3. Create a knowledge base from your files or websites, test retrieval, then build and publish an app or agent.

Sandbox provides a free workspace with 200 trial message credits; that allowance is not stated to recur. Professional is $590/workspace/year with 5,000 message credits/month and three members; Team is $1,590/workspace/year with 10,000 credits/month and fifty members. These are annual totals, not monthly rates. Model-provider charges and quotas still apply. Self-hosting uses a modified Apache 2.0 license; Enterprise commercial authorization and support are separate.

Source: dify.ai/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Dify - The Platform for Production-Ready Agentic Workflows ↗dify.ai
• Dify Pricing - There’s Always a Plan for You ↗dify.ai
• Dify Documentation - Dify Docs ↗docs.dify.ai

Company website

dify.ai: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.