Agents and AutomationSource available

Dify

A visual platform for building and operating AI workflows, agents, and RAG applications.

By LangGenius · reviewed 2026-08-07

What Dify does

Dify combines model-provider management, workflow composition, prompt and agent tools, knowledge bases, APIs, and operational features. It can be self-hosted or used as a managed cloud service, making it an application platform rather than a model runtime.

Pricing model

Source-available self-hosting; paid cloud plans

Position in the stack

Where Dify fits

Dify is an AI application platform above model providers. It combines workflow and agent composition, prompt management, knowledge bases, tools, APIs, and operational controls so teams can publish applications without building every integration from scratch.

Typical deployment flow

A practical sequence for evaluating Dify before making it part of a permanent stack.

  1. STEP 1

    Choose cloud or self-hosting

    Decide whether managed convenience or infrastructure control best fits data, compliance, cost, and operational requirements.

  2. STEP 2

    Configure providers and data

    Connect language, embedding, reranking, and media providers, then define knowledge stores and secrets.

  3. STEP 3

    Build and publish

    Compose workflows or agents, test failure paths and costs, then expose the application through UI or APIs.

Best for

  • Building AI workflows without coding every integration
  • Publishing agent and RAG applications
  • Using local and cloud models behind one application layer

Not the right layer for

  • Direct model inference optimization
  • GPU inventory and cluster scheduling

Capabilities

Web interface
API
Model management
Multi-user
Multi-GPU
Multi-node
RAG
Agents
Image generation
Tool calling

Capabilities refer to the tool's application layer. Hardware and model support can still depend on a connected inference engine.

Deployment and compatibility

Deployment
Self-hostedManaged cloud
Platforms
Linux · Docker · Kubernetes · Managed cloud
Hardware backends
Uses connected model runtime
Install methods
Docker Compose · Kubernetes · Managed cloud

Models and integrations

Model formats
Uses connected model runtime
Common integrations
OllamaOpenAI-compatible APIsCloud model providersVector databases

Strengths

  • Visual workflow builder
  • Broad provider and knowledge integrations
  • Application APIs and operational controls

Limitations

  • Self-hosting introduces several dependent services
  • License conditions differ from a standard OSI license

Planning checklist

Before you choose Dify

Answer these questions with the exact models, hardware, users, and data you expect to operate.

  • Which data and model calls are allowed to leave the private environment?
  • Does the workflow require deterministic branching or autonomous agent behavior?
  • How will secrets, users, quotas, and application versions be managed?
  • Can the team operate the dependent services required by self-hosting?

Dify FAQ

Does Dify replace Ollama or vLLM?

No. Dify builds applications and workflows above model providers. Ollama, vLLM, cloud APIs, and other services can supply the underlying inference.

Can Dify use local models?

Yes, through supported local or OpenAI-compatible providers. Network routing and authentication must allow the Dify deployment to reach those endpoints.

Is self-hosted Dify operationally lightweight?

It is more involved than a desktop application because the platform depends on multiple services. Backups, upgrades, secrets, monitoring, and access control need an operating plan.

Official sources

Use these links to confirm current compatibility and installation requirements.

Last reviewed 2026-08-07

Related tools

Execution evidence

Known working recipes using Dify

Recipes connect hardware, a model artifact, tools, settings, verification, and a reportable result.

Browse all recipes →

No verified recipe is linked to this record yet.

Compatibility estimates remain available in the planner. A recipe appears here only after its exact stack and verification protocol are documented.

Continue your local AI setup