Security at AgentBhau AI

Security for AgentBhau AI is shared between us and the providers we build on. CYBRIX TECHNOLOGIES LLP controls the AgentBhau AI application itself: its features, how it stores and presents your workspace data, how users sign in, and how third-party services are connected. The infrastructure underneath — hosting, networking, the managed database, and the WhatsApp and AI platforms — is operated and secured by those providers under their own terms and documentation.

We do not audit, re-test, or independently certify our providers' internal controls. Their security posture is described in their own documentation, which we link below rather than restating as our own claims.

• Application logic: how AgentBhau AI processes messages, automations, and your workspace data inside the product.
• Team access: access to workspaces is controlled through the application's login and session management, so only signed-in team members reach your workspace.
• API credentials: keys for third-party services (AI models, messaging channels) are stored and used on the server side and are not embedded in public browser code.
• Customer prompts and knowledge: what you upload to your knowledge base and send to AI models is used to answer your customers' questions, and we do not share it with other customers.

• Web hosting: our hosting platform serves the site and runs the server code over standard HTTPS, and is responsible for its infrastructure and network security.
• Database and file storage: the managed cloud provider handles storage infrastructure, including its encryption practices at rest. We rely on the provider's published security documentation rather than making our own storage-engine claims.
• WhatsApp messaging: Meta operates the official WhatsApp Business Cloud API, including its platform-level security and channel compliance, under Meta's developer terms.
• AI and LLM services: model inference runs on third-party providers (such as OpenAI, Google Gemini, and DeepSeek) under their standard API data-use terms. How those providers handle prompts and outputs is governed by their own agreements and documentation.

Official WhatsApp connections use Meta's WhatsApp Business Cloud API through Embedded Signup. The security, encryption, and compliance of that messaging channel are managed by Meta as the platform operator; we integrate with it as an approved developer rather than operating the channel ourselves.

QR-based connections are a separate, non-official mode of linking a number. They are clearly labelled as such in the product, and they do not carry the same platform-managed protections as the official Cloud API.

AI features call model providers through secure server-side API connections. Credentials for these providers live in server-side configuration and are not included in the code delivered to your browser.

Prompts and knowledge-base content are processed by the model providers under their standard API data terms. If you need specifics for a vendor review, contact us and we will confirm which providers are in use and point you to their current documentation.

AgentBhau AI has not completed a third-party independent security audit and does not currently hold formal compliance certifications such as SOC 2 or ISO 27001. We would rather state that plainly than imply controls we have not verified.

For vendor questionnaires, partner reviews, or diligence requests, contact us and we will answer based on our actual practices and the provider documentation described above.

If you find a suspected vulnerability, email [email protected] or [email protected] with a description and steps to reproduce. Please give us reasonable time to investigate and fix before any public disclosure, and do not access or modify other customers' data.

We acknowledge reports as quickly as we can and will let you know what we found and what we changed. There is no bug bounty program at this time.