Artificial intelligence at the service of your QA
Automated software testing, with AI where it genuinely helps.
We build the testing tools we never had. They run on your machines, your data never leaves them.
Our programmes
SAPFX · available today
Our open source SAP test automation ecosystem (SAP GUI, Fiori, API) for
Robot Framework, under the Apache 2.0 licence.
Discover SAPFX →
· Get started in 10 minutes →
CabrIA · R&D programme
An AI model expert in Quality Engineering, fine-tuned locally on your real
test executions and paired with a RAG knowledge base that cites its sources;
100% on-premise, your data stays in.
Discover CabrIA →
CabraFlow · R&D programme
A keyword-driven test orchestrator on top of Playwright: a demo narrated
aloud becomes a business test, replayable natively, as Playwright or as
Robot Framework.
Discover CabraFlow →
Pilot programme
We support a small number of teams on pilot projects, at preferential terms. The frame is always the same: a narrow scope, success criteria agreed up front, and measurements you keep, whether or not the work continues with us. The most frequent questions have their own page: read the FAQ.
What already ships
SAPFX is published as open source: the code is public, the figures are measured, and you can check them before you ever talk to us.
Figures observed at version 0.7.0 (August 2026). The SAPFX page details the method, the GitHub repository lets you check all of it.
Our expertise
- LLM integration with RAG (Retrieval-Augmented Generation)
- Fine-tuning AI models for QA
- Advanced prompt engineering
- AI agents for bug detection and test generation
- Performance testing with AI
- AEO (Answer Engine Optimization) for your visibility on ChatGPT, Claude, Perplexity and Gemini
The technical terms used here are defined in the glossary.
How we work
- No figure we have not measured. Gains depend on your context. We measure them with you on a pilot rather than advertising a generic percentage that fits nobody.
- Nothing heals in silence. When a test finds an element that has changed on its own, the repair is logged. A test that quietly fixes itself stops being a test: it becomes an agreeable witness.
- An AI that cites its sources. Our assistants point back to the documents an answer came from, so you can verify instead of believe.
- Humans decide. Agents plan, write and repair; review stays with the person who knows the business.
- Your data stays in. On-premise deployment on your infrastructure, or on our isolated infrastructure in France.
Security and compliance
- On-premise deployment: AI models installed on your infrastructure or on our isolated infrastructure in France
- No sensitive data transmitted to third parties
- GDPR compliance guaranteed: your data stays with you
- Complete isolation of test environments
Let's talk about your test suite
Tell us where you stand. You will get an opinion, not a brochure.