What is Junyr?
Junyr is a European AI business platform and method for micro-businesses, SMEs and mid-caps. It comes in three forms: Junyr Suite, an AI business management suite (ERP, CRM, projects, shared email and calendar, documents); Junyr Bricks, a custom platform assembled from the same modules around your own processes; and human expertise, with advisory, training and support. Junyr is a brand of Croissance et Transitions, based in Paris, and is pronounced “junior”.
Explore Junyr Suite on junyr.app ↗
What is Junyr’s connected business graph?
It is the knowledge graph at the heart of Junyr. Every email, document, meeting, contact, company and project is linked automatically, with no manual filing. You get a 360° view of each client, and your AI agents work from one source of truth instead of scattered files and inboxes.
How we connect your business data →
Can Junyr run fully on-premises?
Yes. Junyr Suite is hosted in France by default, and a full on-premises deployment, with the entire suite on your own servers, is available as an Enterprise option. Junyr Bricks platforms are deployed on your servers or hosted in France, as you choose. Combined with a local LLM, your data and AI processing never leave your network.
Discuss your architecture ↗
Can I bring my own LLM?
Yes, on every Junyr Suite plan. Your administrator connects any OpenAI-compatible endpoint once, for the whole company: Ollama, vLLM, LM Studio or your own gateway. Open-weight models can run on your own hardware, and you can switch to a better model whenever one comes out. No AI provider is imposed: admins can also pick the model behind each agent layer, your local server or your own Mistral, Gemini or Claude API key.
Local AI and model choice →
How does Junyr protect confidential data?
Three company-wide confidentiality levels decide what any AI model is allowed to read. At the strictest level, no AI touches your data at all. In between, Secure mode replaces names, companies and email addresses with pseudonyms before any cloud LLM call, and Simple mode gives cloud AI full context. With your own local LLM server, eligible AI features run on your hardware instead.
Can we use AI without our data leaving the company?
Yes. Connect your own local LLM server and eligible real-time AI features (writing assist, AI search, smart actions, agent chat, Ask Junyr) run on your hardware. With the on-premises option (Enterprise), the suite and its data stay inside your network too.
Are we locked into Junyr or an AI vendor?
No. AI models are interchangeable, including a switch to local AI. In our advisory engagements, everything we build (code, configurations, agent mandates, prompts and skills) is handed over to you, and you know how to exit each building block before you adopt it. Adopting Junyr Suite is never a condition of our work.
Our written commitments →
How do Junyr AI agents stay under control?
In Junyr Suite, you set each agent’s roles, permissions and a charter it cannot rewrite, and every sensitive action stays approved by a human. In the agentic platforms we build, mandates are also versioned, permissions kept minimal and every action logged. The principle behind the Junyr Method™: the agent proposes, the human commits.
The Junyr Method™ →
Who is Junyr for?
Junyr Suite is built for micro-businesses that want their whole activity connected in one place. Junyr Bricks is for SMEs and mid-caps that need a platform fitted to their own processes. Training and advisory serve leaders and teams at every stage, from AI acculturation to agents in production.
How do I get started?
With a free 30-minute AI maturity audit by video call: you’ll find out where you stand on the five-level Junyr Scale™, what blocks the next level and which first project makes sense. Teams can also start with a one-day in-house AI Acculturation training at €1,900 excl. VAT per group, delivered by a Qualiopi-certified organisation and OPCO-fundable.
The free 30-min audit →Training catalogue →
