D

Dify

agent_platform Website ↗ Website ↗ Verified 2026-11-17

📊 Scores

Coding Ability
7.0
Value
8.5
Flexibility
9.0
China Access
8.0

🚀 Quick Start

Install Steps

  1. 1 Follow repository Quick start: prepare docker/.env and dependencies, then run docker compose up -d in that directory; not a one-command deployment from any folder
  2. 2 Or sign up at cloud.dify.ai to start free
  3. 3 Build agent workflows with drag-and-drop Web UI
  4. 4 Publish as API / MCP Server / embedded page
Register / Get Account ↗
🇨🇳 China Access Note: Self-hostable, but models, plugins, registries and third-party tools may still require external connections; mainland reachability is untested.

💲 Pricing

Plan Price Limits
Self-hosted Community $0 Single workspace; additional Dify License conditions apply; compute/model fees separate Official →
Cloud Sandbox $0 200 message credits (not 200 fixed calls); 1 member, 5 apps, 50 MB knowledge storage Official →
Cloud Professional (annual) $590/year $590/workspace/year; 5,000 monthly message credits, 3 members, 50 apps; not a pay-as-you-go unit rate Official →
Cloud Team (annual) $1590/year $1,590/workspace/year; 10,000 monthly message credits, 50 members, 200 apps Official →

The page defaults to annual billing; paid amounts above are annual totals, not monthly fees. Taxes are extra. Community uses the Dify License based on Apache-2.0 with additional conditions, not unmodified Apache-2.0 or unlimited self-hosting. Enterprise commercial licensing/multiple workspaces/SSO are separate. After model credits run out, bring your own key and pay the provider. Historical editorial scores retained, not newly measured or re-rated; not a fresh performance or cost benchmark.

Dify

Free software/tier listed; check pricing scope and signup terms

Website ↗

✨ Key Features

Code Completion —
Chat ✅
Agent Mode ✅
Multi-file Edit —
MCP Support ✅
Sub-agent ✅
Multi-Agent ✅
agent_teams ✅
task_management ✅
persistent_memory —
skill_sharing ✅
runtime_monitoring ✅
open_source —
self_hosted ✅
visual_builder ✅
rag_support ✅
human_in_the_loop ✅
Context Window 取决于所选模型
supported_runtimes 内置 Agent 引擎
supported_os macOS, Windows, Linux
ide_base Agent 协作平台
source_available ✅

🎯 Best For

Non-technical users rapidly building AI agent appsScenarios requiring RAG + agent workflowsAll-in-one platform from prototype to production

📝 Changelog

2026-10-13 Started tracking version: 1.17.1 (verified via GitHub releases)
2026-05-09 Added to CodePick Agent Collaboration section

🔗 Related Tools

💬 Comments

Last Verified: 2026-11-17 Next Review: 2026-12-17 Confidence: medium Report Error →