Dify Agentic Workflow Platform

Dify is an agentic AI development platform built by LangGenius, Inc. It addresses the gap between AI experiments and production by giving developers and business teams one place to build, connect and operate LLM applications, and it has grown into one of the most-starred open-source AI projects on GitHub.

Users design agents and multi-step workflows in a visual Workflow Studio, ground them in company data through a Knowledge Pipeline that ingests files, websites and drives, and extend them with model, tool, data-source and MCP plugins from a marketplace. Apps can be published as hosted web apps, APIs, embeds or MCP tools, with logs and usage monitoring. Dify is offered as a cloud service, a free self-hosted Community Edition and a self-hosted Enterprise edition used by firms such as Ricoh, Volvo Cars and Maersk.

• Workflow Studio: visual builder for agentic workflows with tool use, logic and human input steps.
• Agents: created by chat or configuration and published as apps, embeds or workflow nodes.
• Knowledge Pipeline: ingests, chunks, indexes and tests retrieval over files, websites and drives.
• Plugin Marketplace: model providers, tools, data sources and MCP integrations.
• Publishing and monitoring: hosted apps, API endpoints and MCP tools with logs, feedback and usage data.
• Enterprise controls: self-hosted or VPC deployment with SSO/SAML, RBAC, audit logs and Kubernetes Helm chart.

IT departments of Vietnamese banks or conglomerates can host Dify on-premises as a shared platform for internal chatbots and agents. E-commerce and retail companies can build customer-service and shopping agents connected to product catalogues. Software firms can use Dify to deliver AI workflows to their own clients.

Shortens the path from prototype to production for RAG assistants and agents with a visual builder instead of custom code. Avoids model lock-in through plugins for many model providers. The Enterprise edition runs inside the customer's VPC with SSO, RBAC and audit logs. A free Sandbox plan and open-source Community Edition allow low-cost evaluation.

Enterprises are standardising on low-code LLM application platforms so that many teams can build agents on shared governance instead of scattered scripts. Open-source, self-hostable tools are popular in Asia, including Japan and Vietnam, where data residency and cost control matter for AI adoption.

LangGenius announced a USD 30 million Series Pre-A round in March 2026. Dify reached 100,000 GitHub stars in June 2025, launched on Azure Marketplace in July 2025 and received a 2025 AWS Partner Award. It holds SOC 2 Type II, ISO 27001 and GDPR compliance, and hosts the IF Con developer event in Tokyo.