Dify
7.3/10Open-source platform for building LLM apps, agents, RAG knowledge pipelines and visual AI workflows, available as cloud SaaS or self-hosted.
Data last updated November 2026 · Our methodology
Overview
Dify, built by LangGenius, Inc., is a platform for developing large language model applications. Teams build chatbots, agents and multi-step workflows on a visual canvas and connect them to knowledge bases. A Knowledge Pipeline ingests files, websites and online drives, then chunks and indexes them for retrieval. You can publish apps as hosted web apps, website embeds, API endpoints or MCP-compatible tools, and logs, annotations and usage data come built in. Dify works with many model providers, including OpenAI, Anthropic, Google, Azure OpenAI, Hugging Face and local models through Ollama. A plugin marketplace adds tools, data sources and MCP integrations. There are three ways to run it. The Community Edition is free to self-host with Docker under an Apache-2.0-derived license, so you pay only for infrastructure and model APIs. Dify Cloud is priced per workspace: a free Sandbox tier, Professional at $590 a year and Team at $1,590 a year, with limits on message credits, apps, members and knowledge storage. Enterprise is quoted by sales and adds SSO/SAML, RBAC, audit logs, multiple workspaces, Kubernetes Helm deployment and 24/7 support. It suits startups building MVPs and enterprises that want to keep data on their own infrastructure. The main limitations come from users: self-hosting takes real DevOps work and can struggle under heavy load, there are fewer connectors than in some rivals, and multi-agent orchestration and observability are basic. You also need a license to remove Dify branding from self-hosted apps.
Our Rating
Who Dify is best for
Good fit
It is an open-source LLM app platform with an API, self-hosting and support for many model providers.
The free self-hosted edition and low-cost cloud plans suit teams prototyping AI MVPs quickly.
The Enterprise edition offers VPC or self-hosted deployment, SSO, RBAC and audit logs, and Dify names large customers such as Maersk and Volvo.
Visual workflows and self-hosting let consultancies build and hand off client AI apps.
The Team plan supports 50 members and 200 apps, which fits internal AI tooling.
Less suited
Self-hosting with Docker or Helm, SSO, RBAC and audit logs support internal deployment.
API endpoints and embeds let software companies ship AI features to customers.
It is affordable, but it still needs technical skills to build apps and integrate them.
Teams can publish chatbots as embeds or web apps, which could be built for clients.
A review describes building a RAG support workflow with human handoff in under an hour.
The free Community Edition and free Sandbox tier lower the cost for individuals.
Pros & Cons
Pros
- The Community Edition is free to self-host with no license fees, so the only costs are infrastructure and model APIs.
- The visual workflow builder lets a team build a RAG support workflow with human handoff in under an hour, according to one review.
- Document ingestion, chunking, embedding and retrieval are handled in a single interface.
- Teams can switch between many model providers per node, including local models via Ollama.
- Apps can be published as web apps, embeds, APIs or MCP tools.
Cons
- Self-hosting needs Docker or Kubernetes skills, and reviewers report scaling problems under heavy workloads.
- Reviewers say Dify has fewer connectors than rivals such as LangFlow.
- Multi-agent orchestration and built-in observability are basic.
- Workflows have no built-in version control, and the canvas can lag on large flows.
- Removing Dify branding from self-hosted apps requires a commercial license.
Key Features
Dify Pricing
Community
- Single workspace
- Self-hosted via Docker
- Dify open-source license
- All core open-source features
Sandbox
- 5 apps
- 1 member
- 30-day log history
- 200 message credits
- 3,000 trigger events
- 10 knowledge requests per minute
- 50 knowledge documents, 50MB storage
- Dify API limited to 5,000 calls per month
Professional
or $590/year (save 17%)
- 50 apps
- 3 members
- Unlimited log history
- 2,000 annotation quota
- Priority document processing
- 20,000 trigger events per month
- 5,000 message credits per month
- 100 knowledge requests per minute
- 500 knowledge documents, 5GB storage
Team
or $1,590/year (save 17%)
- 200 apps
- 50 members
- 5,000 annotation quota
- Unlimited trigger events
- 10,000 message credits per month
- Top-priority document processing
- 1,000 knowledge requests per minute
- 1,000 knowledge documents, 20GB storage
Enterprise
- Negotiated SLAs
- Commercial license
- Multiple workspaces
- SSO/SAML, RBAC, audit logs
- Self-hosted or VPC deployment
- Dedicated CSM and 24/7 support
Frequently Asked Questions
Yes. The self-hosted Community Edition is free, and Dify Cloud has a free Sandbox plan with 200 message credits and 5 apps.
Pricing is per workspace. Professional is $590 a year or about $59 monthly, and Team is $1,590 a year or about $159 monthly. Annual billing saves around 17%.
Only on the Enterprise edition, which adds SSO/SAML, RBAC, audit logs and multiple workspaces.
Dify's website lists SOC 2 Type II, ISO 27001 and GDPR.
It works with OpenAI, Anthropic, Google, Azure OpenAI, Hugging Face and local models through Ollama, among others, and plugins add more providers.
According to community discussion, removing the Dify branding and logo requires a commercial license.
Similar Tools
See all alternatives
Amazon's public cloud platform, with over 200 pay-as-you-go services for compute, storage, databases, networking and AI, used by startups through large enterprises and governments.
Anthropic's organization-wide Claude plan adds chat, Claude Code, Cowork agents, connectors, and enterprise governance such as SCIM, audit logs, and a compliance API.
Fin is Intercom's AI customer service agent. It resolves support conversations across chat, email, voice and social, either inside Intercom or on top of an existing helpdesk.