At Unitalk, trust is the foundation of our platform. We’ve engineered our security and privacy controls from day one to protect your company’s data, ensuring you can innovate with AI confidently and securely.
As a French company, we are bound by and fully committed to the principles of the GDPR. Our entire architecture is built with privacy-by-design at its core.

We adhere to the highest standards of data protection and are committed to maintaining industry-recognized certifications.

RGPD

SOC2

ISO 27001
You have complete control over the data you connect to Unitalk. We provide the tools to ensure your data is handled exactly as you intend.
From upload to response, your data is continuously protected. We employ robust AES-256 encryption for all data at rest and mandatory TLS 1.2+ for data in transit, ensuring security at every step of the process.
You have precise control over which files and data sources form your knowledge bases. Your AI assistants will only access the information you explicitly provide, preventing accidental data exposure and ensuring relevance.
By default, your data is securely hosted on AWS servers in Paris (France) for full GDPR compliance. For maximum control, we offer private cloud and on-premise deployment to meet any data sovereignty requirements.
Securely connect to external applications via our MCP protocol. Every connection is fully encrypted and mutually authenticated, guaranteeing a safe and reliable flow of data between Unitalk and your business-critical tools.
We empower you with leading AI models while ensuring your proprietary data remains private and protected.
We have strict data processing agreements with our AI model providers. Your data is only used to generate a response and is not retained or stored by them after the request is processed.
This is our guarantee. The contents of your knowledge bases, your prompts, and the generated responses are never used to train or improve any third-party AI models. Your data is your intellectual property, period.
Choose the providers and embedding model you want
Integrate seamlessly with the industry’s leading vector databases you already trust, such as Pinecone, Weaviate, or Qdrant.
Secure your Unitalk environment with powerful, easy-to-use access controls.
Assign permissions to your team members. Define roles like Admin, Editor, and Viewer to ensure users only have access to the features and data they need.
Integrate with your existing identity provider (e.g., Google Workspace, Microsoft Azure, Okta) using SAML-based SSO for seamless and secure user authentication.
Your data is never co-mingled. Each AI assistant and knowledge base is logically isolated, guaranteeing that information from one cannot be accessed by another unless explicitly configured by you.
Our commitment to security extends to the very core of our platform and operations.
We conduct regular internal and third-party penetration tests and vulnerability scans to proactively identify and remediate potential threats.
We integrate security into every phase of our development lifecycle, from design and coding to testing and deployment.
We maintain automated backups of your data and have a comprehensive disaster recovery plan to ensure business continuity and data integrity.
Your trust is our most important asset. If you have any questions about our security practices or would like to review our security documentation, please don’t hesitate to reach out to security@unitalk.ai
