Dynamov brings the best AI models together in a single platform: your teams become more productive, you match every use case to the right cost, and you track ROI on dedicated dashboards — with European sovereignty and GDPR & EU AI Act compliance as standard.
The platform
Dynamov covers the whole chain: chatting, building business assistants, automating tasks, and governing all of it. Two ways to use AI, depending on the need.
The default experience, usable with no AI knowledge at all. Users pick a complexity level — Simple, Standard, Complex, Very complex — instead of a model. Each level shows its sovereignty flag 🇪🇺/🇺🇸 and its cost ratio.
Specialised AI agents for a precise role (Advisor, Financial analyst, Email writer…), built without code by your administrators. Each agent combines a model, internal knowledge, a framed behaviour and governed tools.
From sovereign European models to premium international ones. The right model for the right use, without switching tools.
Assemble business assistants in a few clicks, across four dimensions. Continuously improved from user feedback.
Scheduled agents, multi-step workflows, mailboxes handled automatically. AI does the work, a human approves.
Connect your internal documents, vectorised in France. The AI answers from your real documents and cites its sources.
A-priori anonymisation of sensitive data and a-posteriori scanning by a sovereign model. GDPR & EU AI Act compliant.
ROI, adoption, security, footprint, costs: six ready-to-use dashboards for leadership and managers.
Why Dynamov
Nearly 30 AI models including GLM-5.2, Claude Opus 4.8, GPT-5, Qwen3, Deepseek v4, Gemma — consistently state of the art. Pick the right model for each use case. One platform, no technical trade-off forced on your teams.
Models hosted in European datacenters (Qwen3 235B and GLM-5.2 in Finland on Nebius, Mistral in France), vector data stored in France. Configure full sovereignty on sensitive use cases, model by model.
A-priori protection before every exchange, a-posteriori scanning of conversations, real-time alerts, an immutable audit log and GDPR/EU AI Act compliance measured continuously. Your data doesn't leak, and neither do your costs.
2 months free to measure the gains, ROI dashboards built for the executive committee, and a simple guarantee: no measurable gains, you stop at no cost. Activated in your organisation with no infrastructure project.
Chat Mode
Your employees don't have to choose between Claude, GPT, Mistral or Qwen. They pick a level suited to their question. Behind each level sits a system agent fully configured by your organisation.
Tailored agents
Every Dynamov agent is set up across four dimensions. In a few clicks. User feedback (👍/👎) feeds continuous improvement, measured by your administrators.
Choose from nearly 30 models — Claude, GPT, Mistral, Qwen, GLM, Gemma — based on the power / cost / sovereignty trade-off you want. A panel shows power, cost and region for each choice.
Connect your document bases (vector RAG) hosted in France, with 3 visibility levels. Models only see what you allow, and cite their sources.
Define the role, tone, language, guardrails and systematic refusals via a system prompt. Your agent stays true to your brand and values.
Enable governed tools: document search, web reading, calculations, data analysis and Gmail. For emails, the agent prepares a draft; you review and send.
Beyond chat
Import a bank statement, an Excel file or a CSV: the agent computes totals and breakdowns and plots charts in the chat. Raw data never leaves Europe — only the schema and aggregated results are shared.
🇪🇺 Categorised by a European modelScreenshot, photo of a document, diagram, table: the AI describes, extracts text or answers questions about it. Native vision for capable models, otherwise back-up transcription by a European model.
Up to 5 files per messageIllustrate a presentation or prototype an idea from a description. Sovereign generation via FLUX served from a European cluster, or premium via OpenAI GPT Image 1.5. Download or copy in one click.
FLUX (EU) or OpenAIFor capable models, real-time internet search via the provider's API, and automatic reading of cited web pages. The agent works with up-to-date information when you allow it.
Enabled per agentAgentic automations
Beyond conversation, Dynamov runs recurring tasks: on a schedule, in a chain, or in response to an event. Always with a human in the loop for sensitive actions.
Strategic insight
Six built-in dashboards, with no third-party tool to set up. Real previews of the admin console.
WAU/MAU, cohort retention, onboarding, top active users. Spot champions and blockers.
Hours saved, consolidated ROI in euros, satisfaction, trajectory. The dashboard to present to leadership.
Anonymisation, detected leaks, exceptional accesses, EU AI Act compliance score, immutable audit log.
Cost × power × sovereignty chart. Four quadrants to find the sweet spot and avoid waste.
CO₂, water and energy footprint per request (EcoLogits), savings vs a 100% US usage, optimisation recommendations.
Credits consumed, top models, breakdown by provider, credits ⇄ tokens toggle. Caps and alerts to prevent any drift.
Security, compliance & governance
Data protection can't rely on a single filter. Dynamov combines upstream defence, downstream analysis, and fine-grained access governance — all measured and auditable.
Compliance score self-assessed against 10 rules (GDPR Art. 25/30/32/33/35, AI Act Art. 13/14/15, EU-first sovereignty, CNIL), with per-rule status and real-time alerts.
Automatic detection of names, emails, phone numbers, addresses, IBANs, company IDs, amounts and company names, replaced with placeholders before being sent to the model.
Four roles (Super Admin, Org Admin, Creator, Member). Your organisation is strictly isolated: your members, agents and conversations are never visible elsewhere.
Conversations are strictly personal. Admins only see metadata; access to content requires an exceptional, approved and logged request.
Every sensitive action — access, sharing, role change, reset — is recorded immutably and can be reviewed.
Your business data, vectorised in France, is never shared with foreign models or used to train them. Three visibility levels per base.
Hybrid sovereignty
Sovereignty isn't binary. On Dynamov, each use case chooses its model, its hosting and its confidentiality regime.
Budget control
From economical to premium, cost varies by a factor of 100. A centralised platform lets you arbitrate in real time by use case : a light model for simple tasks, a powerful one for critical stakes. Your AI costs stay under control, by design.
Pricing
€199 excl. VAT per month for your entire organisation, with an included pool of credits to share across your teams. Top up only if usage grows. And before you pay, 2 fully free months to measure the gains.
For your whole organisation
For larger deployments: contractual SLA, dedicated support, reinforced sovereign deployment and EU AI Act compliance support — available within the same offer, negotiated at signing.
Talk to an expertThe company
Dynamov is a simplified joint-stock company (SAS) registered in France, based in Versailles. The platform is built and operated in-house, then opened to a pilot partner programme for wider deployment.
Grégoire Semelet
Founder
Behind Dynamov: Grégoire Semelet, an IT engineer with 18 years of digital transformation experience with major French accounts — Orange, AXA, Air France, Accor, Bouygues Construction.
He also founded MaCertif, a training organisation that has supported more than 2,200 professionals across 25 countries in project management, digital and responsible generative AI. For him, sovereignty, GDPR and the EU AI Act aren't checkboxes, but the foundations of high-performing, responsible AI.
“Agentic AI shouldn't be a luxury reserved for the American giants. Our ambition with Dynamov: to make it powerful, sovereign and accessible to every European SME and mid-market company — without ever trading off productivity against security.”
Frequently asked questions
Dynamov is a European, hybrid and sovereign agentic AI platform for SMEs and mid-market companies. In a single interface it brings together nearly 30 leading AI models (GLM-5.2, Claude Opus 4.8, GPT-5, Mistral Large, Qwen3 235B, Deepseek V4, Gemma 4 31B), no-code agent creation, automations, knowledge bases, data protection and management dashboards — with a sovereignty level configurable per use case. The platform is published by Dynamov, a French SAS based in Versailles.
Dynamov is fully free for 2 months so you can concretely measure the gains (productivity, savings, quality). Beyond that, a flat fee of €199 excl. VAT per month covers the whole organisation, with an included pool of credits to share across users. The price doesn't change if you hire; you only top up if usage grows. If no gains are observed, you stop at no cost.
Yes. The vector knowledge bases are hosted in France. Models can be chosen by location: European datacenter (Qwen3 and GLM-5.2 hosted in Finland on Nebius, Mistral in France) for full sovereignty, or best-in-class non-European models (Claude, GPT) for non-sensitive use cases. A 100% sovereign mode can be switched on.
Dynamov combines two layers of protection: an a-priori filter that automatically anonymises sensitive data (names, emails, IBANs, company IDs, client identifiers, trade secrets) before sending it to the model, and an a-posteriori cross-cutting scan that analyses conversation history via a sovereign model to detect residual leaks or risky behaviour.
Yes. Dynamov is privacy-first by design: models hosted in European datacenters, data stored in France, a built-in EU AI Act compliance dashboard, a GDPR score self-assessed against 10 rules (GDPR Art. 25/30/32/33/35, AI Act Art. 13/14/15, EU-first sovereignty, CNIL), an immutable audit log and real-time alerts.
Chat Mode lets every employee choose a complexity level (Simple, Standard, Complex, Very complex) instead of a model. Each level shows its sovereignty flag and cost ratio. A « + » escalation feature replays a question at the higher level and compares the two answers side by side: you only pay for the higher level when it's genuinely useful.
Yes. Each agent is built in a few clicks with no code, across four dimensions: choosing the LLM (brain), connecting business knowledge bases (knowledge), defining behaviour (tone, guardrails, refusals), and enabling governed tools — document search, web reading, calculations, data analysis, Gmail mailbox. For emails, the agent prepares drafts: a human reviews and sends.
Yes. An agent equipped with data tools lets you import a bank statement, an Excel export or a CSV and get tables and charts directly in the chat. The analysis is sovereign: raw data is never sent to the model in bulk (only the schema and aggregated results are shared), and categorisation relies on a European model. A model such as Qwen3-235B, hosted in Europe, is recommended.
Yes. Dynamov offers agentic automations: scheduled agents (for example a news digest every morning), multi-step workflows chaining several agents with context passing, and automatic mailbox handling (IMAP or Gmail) — the agent learns your organisation's email types and prepares reply drafts with the right attachments. Never auto-sent: a human reviews and sends.
Dynamov includes ready-to-use dashboards: consumption and credits, adoption (WAU/MAU, retention), strategic vision for the executive committee (hours saved, ROI in euros), model comparison (cost × power × sovereignty), security (GDPR/EU AI Act, anonymisation, audit) and footprint (CO₂, water, energy via EcoLogits). Everything is built in, with no third-party tool to set up.
Get started
30 minutes with a Dynamov expert to assess your top use case, quantify the expected ROI and kick off a pilot.