Dify

Dify is heading to ATxSG 2026 in Singapore! 🚀 From May 20 to 22, come visit us at Booth 3K2-1 in Halls 3 at Singapore EXPO. Stop by to see our latest live demos, chat with our team about enterprise use cases, and explore how we can work together. Slots for 1 on 1 sessions are limited, so make sure to book your meeting with us below! 👉 Secure your slot here: https://lnkd.in/g-MKF4MJ #ATxSG2026 #Dify #AI #LLMOps #AIAgents

🔥 Dify v1.14.1 is out. It’s been a while since the last release, and during that time we’ve been listening closely to community feedback, refining details, fixing rough edges, and rethinking how workflows should actually work in real teams. This release is less about adding more things onto the canvas, and more about making workflows collaborative, reusable, and easier to integrate into existing systems. What ships: - Collaboration Mode: Multiple teammates can now edit the same workflow in real time, with comments, @ mentions, email notifications, and live collaboration helping business, product, and engineering align directly on the canvas itself. - Copy & paste nodes anywhere: Every node (except user input) is now reusable across workflows and even across Dify instances. Proven workflow logic no longer needs to be rebuilt from scratch every time. - HITL Service API: Human input steps can now integrate with enterprise IMs, or internal systems through API access, instead of being limited to the Dify Console. - Stability Improvements: A focused round of fixes across MCP, OAuth flows, plugin upgrades, Marketplace interactions, and overall runtime stability. Read the blog: https://lnkd.in/gmG3UqjB Full release notes: https://lnkd.in/g46ST7q7 Docs: https://lnkd.in/gQRFjBWk

Final call! Join us at AWS Builder Loft in San Francisco tomorrow, May 7, for an evening with Amazon Web Services (AWS), Dify, and TiDB, powered by PingCAP on what it really takes to move AI agents and GenAI apps from prototype to production. If you’re working on AI agents, production GenAI workflows, or enterprise AI apps, join us to connect with other builders and learn what it takes to ship reliable, scalable AI systems in the real world. See you there! 📍 AWS Builder Loft 525 Market St, 2nd Floor, San Francisco, CA 📅 Thursday, May 7, 2026 🕔 5:00 PM – 8:30 PM Register here: https://luma.com/v9no5ytc

LangGenius K.K.’s official partner Open Network System (ONS) is at AI EXPO KOREA on May 6-8, showcasing and introducing Dify! If you're attending the expo, don’t miss the chance to stop by the booth, see live demos about Dify, and chat with the team. 🚀 #AIEXPOKorea #Dify #AI #ONS

Join us at AWS Builder Loft in San Francisco on May 7 for an evening with Amazon Web Services (AWS), Dify, and TiDB, powered by PingCAP on what it really takes to move AI agents and GenAI apps from prototype to production. In collaboration with TiDB, powered by PingCAP, we’ll be bringing together Bay Area AI app developers, builders, and teams working on scalable GenAI systems for practical talks, demos, and networking. If you’re building AI agents, production GenAI workflows, or enterprise AI applications, this is a great chance to connect with other builders and learn what it takes to ship systems that are reliable, scalable, and ready for real-world use. 📍 AWS Builder Loft 525 Market St, 2nd Floor, San Francisco, CA 📅 Thursday, May 7, 2026 🕔 5:00 PM – 8:30 PM Register here: https://luma.com/v9no5ytc

On 4/16, Alvin Wu (Head of Middle East & Taiwan) and Kelvin Tsim (APAC Solution Architect) from Dify went hands-on with New York University Abu Dhabi MBA students, faculty, and NYU Stern School of Business alumni. One question drove the session: what does an AI agent look like when it's shipping work every day? Three things we actually showed (not just talked about): • A scheduled financial analyst agent — runs hourly, aggregates web topics, passes them to a node set up as a senior Abu Dhabi analyst, and routes output through email-based human approval before it ships. Every node logged, every answer traceable. • Dify’s new Human-in-the-Loop node — the missing piece most enterprises need to move past PoC. Critical decisions pause for review; everything else keeps running. • The Knowledge Pipeline — ETL preprocessing + LM-driven Q&A generation + per-chunk citations. The gap between “works in a demo” and “holds up in production.” Alvin also unpacked why ~80% of AI projects never escape pilot — and showed a manufacturing client that built 200+ apps, kept 60 in production after a month, and cut analysis time from 8h to 3h. Thanks to the AI and Tech Society at Stern at NYUAD MBA organizing team! Special thanks to Treedom Zhang and NYU Stern School of Business for making this session happen.

You've built your AI agent on Dify. The drag-and-drop workflow is smooth, the RAG pipeline is solid. But how do you know it actually holds up when real users start pushing it in unexpected directions? That's where ArkSim comes in. Dify makes it fast to build and deploy production-ready agentic workflows. ArkSim closes the gap between "it works in testing" and "it works in the wild" automatically simulating thousands of real-world conversations to surface failures before your users do. Build fast. Ship with confidence. Try Arksim on your next Dify project: https://lnkd.in/eQUjS76A #Dify #AIAgent #AgentEval #ArklexAI #OpenSource

Do we really need to choose between automation and agentic AI? The debate is often framed as a binary choice. In practice, it rarely is. Over the last 2–3 years, we have created real value at Maersk through what I think of as hybrid architecture — or intelligent automation. The model is simple: combine deterministic processes with targeted AI/LLM capabilities in the same orchestration framework, and use each where it performs best. A good example is email response automation. AI is excellent at intent detection and information extraction from incoming emails. Deterministic workflows then handle structured data retrieval, business rule execution, and predictable system updates. That combination delivers something much more useful than either approach on its own: greater speed, stronger accuracy, and better control. A major part of making this work at scale has been Dify.ai. It has been a key enabler by providing a common platform for rapid development and production rollout — with less code, reusable components, and a practical way to blend AI capabilities with deterministic workflows. With the right platform and some innovative thinking, we have been able to move from idea to production in a matter of weeks for high-value, real-world business use cases. Agentic AI is powerful, but in many enterprise settings the real breakthrough comes from combining intelligence with control. #AI #AgenticAI #Automation #IntelligentAutomation #EnterpriseAI #DifyAI #LLM #WorkflowAutomation #DigitalTransformation #BusinessTransformation