TL;DR: Dify is an open-source platform for building AI chatbots, agents, and workflows on your own data. Its cloud Sandbox plan is genuinely free with 200 message credits, 5 apps, and 50 knowledge documents — and if you self-host the Community Edition, the software itself costs nothing at all. The catch is simple: the free cloud credits do not refill monthly, and model usage beyond them is on you.
What does Dify actually do?
Dify lets you build AI applications without starting from raw API calls. You get a visual studio for chatbots, text generators, agents, chatflows, and workflows, plus a knowledge base (RAG) system: upload documents, web pages, or Notion content, and your app can answer from that material instead of guessing. Apps can be published as web apps or used through an API, and a marketplace adds models, tools, and templates.
It supports models from OpenAI, Anthropic, Google Gemini, and xAI — the free credits let you try them without configuring your own API key first.
What is actually free on Dify?
According to Dify’s official pricing page, checked October 22, 2026, the Sandbox plan is free and includes:
- 200 message credits to try models from OpenAI, Anthropic, Gemini, xAI, and Tongyi
- 1 team workspace and 1 team member
- 5 apps
- 50 knowledge documents and 50MB of knowledge data storage
- 3,000 trigger events and 5,000 API rate limit per month
- 30 days of log history and standard document processing
Two details matter. First, Dify’s own education page describes the same 200 credits as a total allowance with no monthly reset — once they are used up, you switch to your own model provider API key and pay that provider directly. Second, the self-hosted Community Edition is free open-source software you run on your own infrastructure. That removes Dify’s plan limits, but not your server or model costs.
What do the paid plans add?
Professional is $59 per workspace per month ($590 a year) with 5,000 message credits a month, 3 team members, and 50 apps. Team is $159 per workspace per month ($1,590 a year) with 10,000 credits a month, 50 members, and 200 apps. Enterprise is custom. In other words, paying Dify is mostly about hosted credits, team size, and scale — not about unlocking the basic builder.
Pros and cons
Pros
- Real builder on the free plan — chatbots, agents, workflows, and RAG included
- 200 credits cover several model providers, so you can compare models in one place
- Free self-hosted Community Edition if you outgrow the cloud free plan
- Publish as a web app or through an API
Cons
- 200 credits is a trial-sized allowance, not a monthly budget
- Only 1 team member and 5 apps on Sandbox
- Self-hosting means you own setup, updates, and uptime
- Model costs are separate once you bring your own key
Who is Dify best for?
Dify’s free plan is best for developers, technical founders, and curious builders who want to prototype a support bot, a document Q&A app, or an agent workflow before spending anything. Non-technical users can use the visual builder, but self-hosting and API keys reward some technical comfort.
FAQ
Is Dify really free?
Yes, in two ways: the cloud Sandbox plan is free with 200 message credits, and the Community Edition is free to self-host. Model usage beyond the included credits is paid to the model provider.
Do Dify’s 200 free credits reset every month?
Dify describes the Sandbox allowance as 200 credits in total, with no monthly reset stated on its education page. Treat it as a starter allowance, not a recurring budget.
Can I use Dify without coding?
The visual studio is built for low-code work — templates, drag-and-drop workflows, and knowledge uploads. Publishing, APIs, and self-hosting are easier with technical help.
What happens when my free credits run out?
You can add your own model provider API key and keep using Dify, upgrade to Professional for 5,000 credits a month, or self-host the Community Edition.
Research-based review — no hands-on testing claims. Free-tier details checked October 22, 2026 against Dify’s official pricing and education pages. Plans can change; check Dify before you rely on a limit.