● Beta — KubeEra is in active development. Profiles are live; AI-perception measurement is rolling out.
KubeEra
D

Dify

AI Agents
Open source · NOASSERTION

Dify is a platform for building and running agentic workflows and LLM-backed applications—chatbots, RAG pipelines, multi-step agents—through a visual builder plus an underlying API/SDK layer.

updated 2026-07-01

What it is

Dify is a platform for building and running agentic workflows and LLM-backed applications—chatbots, RAG pipelines, multi-step agents—through a visual builder plus an underlying API/SDK layer. It bundles prompt orchestration, retrieval, tool-calling, and observability into one deployable stack rather than making you wire together separate frameworks (LangChain for orchestration, a separate vector DB UI, a separate eval tool, etc.). It's positioned as "production-ready," meaning the pitch is that you can go from prototype to deployed agent without re-platforming.

Who builds it and why

1,373 contributors and 147k GitHub stars indicate a large, active open-source community, not a single-vendor side project. The scale of contribution suggests real usage driving the codebase forward, not just marketing-driven star farming. Language is TypeScript, consistent with a product that leans heavily on a polished web UI as the primary interface, not just a backend library. Founding date and CNCF status are not publicly available—this is not a CNCF project, so don't expect CNCF governance, security audits, or graduation milestones as a trust signal here.

Production readiness signal

Commit velocity (last commit essentially current) and release cadence (1.15.0, well past initial 1.0) point to active maintenance and a maturing API surface. That's a good sign for teams worried about abandonment.

The concerning signal is the license: NOASSERTION. GitHub couldn't detect a standard OSI license, or the repo uses a custom/dual license without clear machine-readable terms. Before you deploy this in a commercial product, get your legal team to read the actual LICENSE file—don't assume MIT/Apache-style permissiveness. This is the single most important thing to verify before adoption, full stop.

Beyond licensing: "production-ready" is a vendor claim, not a certification. Validate it yourself—load test the retrieval pipeline, check how it handles LLM provider failover, and confirm your observability/tracing needs are actually met by the built-in tooling versus needing to bolt on your own.

Who should use this

  • Teams that want a single self-hostable platform for LLM app development instead of assembling LangChain + a vector store UI + a separate eval framework.
  • Product teams who need a visual workflow builder so non-ML engineers (or less code-heavy teams) can iterate on agent logic without redeploying code.
  • Organizations already comfortable running open-source infra themselves who want to avoid vendor lock-in to a single closed LLM app platform.

Who should NOT use this

  • Anyone who hasn't resolved the license ambiguity. If your legal/compliance team can't get comfortable with a NOASSERTION license, this is a hard no regardless of technical fit.
  • Teams needing CNCF-grade governance guarantees (vendor-neutral steering, documented security process, graduation criteria). That signal doesn't exist here.
  • Heavy enterprises with existing investment in LangChain/LlamaIndex + custom orchestration who don't want to migrate workflow logic into Dify's visual/DSL format—rewriting working pipelines to fit a new platform's abstractions is real cost.
  • Teams that need extremely fine-grained control over orchestration logic in code (complex branching, custom retry/backoff, non-standard tool chains)—visual builders trade flexibility for speed, and you may hit walls.

Alternatives

  • LangChain / LangGraph — code-first orchestration framework with more flexibility but more assembly required; no built-in UI.
  • Flowise — similar visual low-code LLM app builder, smaller community, lighter weight.
  • LangSmith / LangFuse (paired with a framework) — if you just need observability/eval rather than a full app platform, these are narrower, more composable options.

Pricing

Core platform is open source (self-hostable). Check the actual LICENSE file for terms given the NOASSERTION status before assuming standard OSS obligations apply. Dify also offers a hosted cloud SaaS tier with its own pricing—not covered by the open-source terms. Confirm which tier you're evaluating before comparing costs against alternatives.

Frequently asked

What is Dify?+
Dify is a platform for building and running agentic workflows and LLM-backed applications—chatbots, RAG pipelines, multi-step agents—through a visual builder plus an underlying API/SDK layer.
Who builds Dify?+
1,373 contributors and 147k GitHub stars indicate a large, active open-source community, not a single-vendor side project. The scale of contribution suggests real usage driving the codebase forward, not just marketing-driven star farming.
Is Dify production ready?+
Commit velocity (last commit essentially current) and release cadence (1.15.0, well past initial 1.0) point to active maintenance and a maturing API surface. That's a good sign for teams worried about abandonment. The concerning signal is the license: NOASSERTION.
Who should use Dify?+
Teams that want a single self-hostable platform for LLM app development instead of assembling LangChain + a vector store UI + a separate eval framework. Product teams who need a visual workflow builder so non-ML engineers (or less code-heavy teams) can iterate on agent logic without redeploying code.
Who should not use Dify?+
Anyone who hasn't resolved the license ambiguity. If your legal/compliance team can't get comfortable with a NOASSERTION license, this is a hard no regardless of technical fit. Teams needing CNCF-grade governance guarantees (vendor-neutral steering, documented security process, graduation criteria).
What are the alternatives to Dify?+
LangChain / LangGraph — code-first orchestration framework with more flexibility but more assembly required; no built-in UI. Flowise — similar visual low-code LLM app builder, smaller community, lighter weight.
How much does Dify cost?+
Core platform is open source (self-hostable). Check the actual LICENSE file for terms given the NOASSERTION status before assuming standard OSS obligations apply. Dify also offers a hosted cloud SaaS tier with its own pricing—not covered by the open-source terms.