Dify

AI app and workflow platform with model connections, knowledge pipelines, tools, and cloud or self-hosted deployment.

Build an agent workflow, connect a knowledge base, and test it before choosing the cloud or self-hosted deployment route.

Access and pricing: Cloud Sandbox is free with limited trial message credits. Community self-hosting has licence conditions and requires infrastructure and model access.

What it does well

The official pricing page shows Dify offering a free Cloud Sandbox, paid cloud plans, a free self-deployed Community edition and a custom-priced Enterprise option. The paid plans scale team members, apps, monthly message credits and knowledge storage, which gives a clear upgrade path if a prototype grows.

Getting started in 10 minutes

Based on documentation, create a Sandbox workspace, build one small app, upload a few documents to a knowledge base and test questions against them. Decide only afterwards whether cloud or self-hosting fits. Community self-hosting runs with Docker in a single workspace.

Limits to check before you commit

Sandbox lists 1 team member, 5 apps, 200 message credits, 50 knowledge documents and 50 MB of storage. The page does not say whether Sandbox credits renew, while paid plans specify monthly credits. Professional is $590 per workspace per year and Team is $1,590 per workspace per year. Enterprise adds SSO, multiple workspaces and negotiated SLAs through sales.