HIVEMIND

HIVEMIND is rethinking the future of software engineering through AI collaboration, openness and responsibility.

Discover more about the project:
👉 hivemind-project.eu

@dfki.bsky.social @australo.bsky.social @ua.es @upc.edu @qubelfastofficial.bsky.social

Software that learns from itself. HIVEMIND’s AI can detect issues, suggest solutions and optimise performance, paving the way for self-repairing systems built on trust and transparency.

Reliable software means trustworthy software. HIVEMIND integrates automated verification and testing tools that ensure every line of code performs as intended, safely and consistently.

Better software starts with better design. HIVEMIND helps define clear requirements, reduce ambiguity and align system architecture from the start, improving collaboration across the development process.

Imagine an AI that codes alongside you. HIVEMIND’s multi-agent system assists developers in writing, testing and improving code, making software development faster, smarter and more reliable.

HIVEMIND promotes collaboration between humans and AI through a shared, open-source ecosystem. Our framework encourages responsible innovation and transparency, ensuring AI supports people, not replaces them.

#HIVEMIND develops an open, scalable and responsible #AI framework that supports developers and transforms how software is created and maintained.

Learn more: hivemind-project.eu

#SoftwareEngineering #HorizonEurope #OpenSource

TIGA Healthcare Technologies leads a HIVEMIND use case developing secure AI software for organ auto-contouring in medical imaging. The system will track compliance, align verification, & optimise deployment to achieve 95% data protection accuracy and 95% AI-driven deployment success: bit.ly/3KTjRB2

💡 New #HIVEMIND paper by Quim Motger, Marc Oriol, Max Tiessler, Xavier Franch & Jordi Marco of the @upc.edu

Fine-grained emotion extraction from app reviews offers new ways to capture user needs. LLMs show promise, but balancing efficiency with trust remains key.

👉 hivemind-project.eu/resources/

New #HIVEMIND research. Marc Oriol, Quim Motger, Jordi Marco & Xavier Franch (@upc.edu) show how Multi-Agent Debate among LLMs boosts requirements classification, reduces bias and opens new possibilities for collective reasoning in RE.

👉 hivemind-project.eu/resources/

In disaster response, speed and accuracy save lives. #HIVEMIND ’s second use case in Türkiye, led by HAVELSAN, applies AI to coordinate emergency operations. Impact: 40% faster design, 70% accurate task estimation, 80% coding compliance.

Read more about our use cases: hivemind-project.eu/use-cases/

#HIVEMIND’s first use case starts in Romania, led by SIMAVI. AI agents help capture requirements, align systems, and cut redundant coding work.

Impact: 70% requirement completeness, 60% faster development, 80% coding advice accuracy.

👉 Read more about our use cases: hivemind-project.eu/use-cases/

#HIVEMIND has kicked off Work Package 8: Integration & Validation of the Framework.

Led by SIMAVI, the work package focuses on:
🔹 integration methodology & QA
🔹 framework validation
🔹 cybersecurity checks
🔹 final refinement

More about us👉 hivemind-project.eu

👉 Back in August, we released our short explainer video about #HIVEMIND, showing how #AI agents work together and where the system is being applied in practice.

If you haven’t seen it yet, now’s the time: hivemind-project.eu/hivemind-exp...

#SoftwareDevelopment

Behind #HIVEMIND are the people who make it happen.
In our first months, Work Package Leaders shared their role, expertise & the challenges ahead.

All interviews are available in our News section 👉 hivemind-project.eu/news/

#SoftwareEngineering

Software dev is stuck in bottlenecks: endless fixes, legacy code & fragmented tools. Most #AI assistants add friction. #HIVEMIND transforms AI into a true collaborator, streamlining the entire SDLC with multi-agent support.

More: hivemind-project.eu/about/

Even in the middle of August, the #HIVEMIND team is advancing towards delivering innovative solutions for software engineering. This week, we reviewed the work packages, assessed potential risks, and mapped out the activities for the second half of the year.

👉 hivemind-project.eu

We’ve created a short explainer video that presents how #HIVEMIND works. It introduces the core concept, the role of #AI agents, and the practical use cases where the system is being applied.

Watch the video: hivemind-project.eu/hivemind-exp...

Thanks @australo.bsky.social for your work!

Missed our first newsletter? No problem, you can find it in the Resources section on our website: hivemind-project.eu/resources/

Get future updates straight to your inbox: a06c99a4.sibforms.com/serve/MUIFAM...

We’ve put together a short, clear presentation on what #HIVEMIND is, and why it matters.

Discover how our human-centred, multi-agent AI framework is reshaping software engineering with responsibility and precision.

📑 hivemind-project.eu/resources/

🧠 #HIVEMIND is putting #AI to work across 5 real-world use cases:
• Citizen Engagement
• Disaster Response
• Healthcare
• Automated Mobility
• Manufacturing

Smart agents cut dev time, boost compliance & improve code quality👉 hivemind-project.eu/use-cases/

🧠 It’s time to sync minds! On July 2–3, 2025, the #HIVEMIND consortium gathers in Belfast for our plenary meeting, hosted by Queen’s University Belfast.

We’re reviewing progress, sharing WP updates, and exchanging ideas to boost collaboration and collective intelligence.

Why open-source? Because rethinking software shouldn’t be locked behind closed doors. HIVEMIND is committed to making its tools publicly available to benefit developers, industry, and society.

👉Learn more about our approach: hivemind-project.eu

And yes, it goes beyond the code. HIVEMIND supports the entire software lifecycle—from smart requirements to updates and maintenance across architectures. Tested in 5 real-world environments. Built to last.

This isn’t about one-size-fits-all AI. Each HIVEMIND agent is tailored to the job at hand: planning, testing, maintenance, helping reduce rework and support better decisions, early on.

HIVEMIND brings together experts to redesign software development, by making it collaborative between humans and AI agents that understand specific roles and tasks. Think of it as a smarter, role-based assistant for every step of the process.
#SoftwareEngineering

Software powers nearly everything—from hospitals to mobility systems. But the way we build it? Still too rigid, slow, and hard to adapt. That’s where #HIVEMIND comes in.
Let’s talk about AI, collaboration, and a new way forward for software.👇

#HumanCentricAI

Get the latest from HIVEMIND.

Subscribe to our biannual newsletter for a concise update on key developments and insights, delivered directly to your inbox. Stay informed and ahead of the curve.

👉 Sign up here: bit.ly/4lX1NEn

🇨🇭 In #HIVEMIND, SUPSI (via IDSIA) develops the core multi-agent framework for responsible #AI. From designing agent interaction & APIs to customising LLM-based agents, SUPSI shapes how AI systems collaborate, adapt & stay reliable.

👉 bit.ly/3EpABNu

13 partners. 8 countries. 5 real-world use cases.
#HIVEMIND isn’t about AI hype, it’s about making software engineering smarter, together.

🧠 Learn more about our approach: hivemind-project.eu