Scope
This policy covers the public website at durbash.com, including its live AI voice and chat previews, product pages, and links for contacting Durbash.
A future customer workspace or a paid Durbash service may involve additional data and will be governed by the service agreement, order form, or data-processing terms presented for that service.
Operator identity
The operator and published legal identity for this public website is Durbash.com.
Durbash.comBuilding No. 23, Subhash Nagar, Saharanpur, Uttar Pradesh 247001, IndiaPhone: +91 85298 97855Email: [email protected]Privacy questions and requests may be sent using these details. Email is the preferred route for requests that require a written response.
Information processed
Information you choose to provide
- Chat preview: the messages you type and the recent conversation turns needed to answer you.
- Voice preview: microphone audio after you press the microphone button, live transcripts, session-negotiation data, and limited product-guide context such as your business type or communication goal.
- Contact: the email address and information you choose to send when a Durbash link opens your email application or an external booking destination.
Technical information
The site and its infrastructure process ordinary connection data such as IP address, user-agent information, request headers, requested URLs, timestamps, and error metadata. The Durbash AI endpoints use this information to secure requests and apply rate limits.
For AI requests, Durbash derives a one-way, pseudonymous safety identifier from connection information. OpenAI receives that opaque value in the safety-identifier field rather than the raw address. Separately, a live voice session uses SDP negotiation and a direct WebRTC connection, so OpenAI may receive the IP address and other network metadata needed to establish the realtime transport. Cloudflare also processes network information while delivering and protecting the site.
How the live AI previews work
| Experience | What is sent | Current Durbash behavior |
|---|---|---|
| Streaming chat | Your validated recent conversation and a pseudonymous safety identifier are sent to OpenAI's Responses API. | The response streams back to your browser. The request sets store: false, and the public-site code does not write the conversation to a Durbash database. |
| Live voice | Microphone audio, SDP session-negotiation data, WebRTC network metadata, transcripts, and limited product-guide context are processed by OpenAI after you start. | The Durbash server brokers setup, then your browser exchanges realtime media with OpenAI. Audio and transcripts are not written to a Durbash database by the current public-site code. |
Public preview tools can explain Durbash and move through a product-guide workflow. They cannot book appointments, send customer messages, or change business records.
How information is used
- Provide the website, AI previews, transcripts, and requested responses.
- Protect the service, enforce same-origin rules, prevent abuse, and limit request volume.
- Diagnose failures using bounded operational metadata rather than deliberately logging conversation content.
- Respond when you choose to contact the Durbash team.
- Comply with applicable law and protect the rights, safety, and integrity of users, Durbash, and service providers.
Service providers and external destinations
- OpenAI processes text, audio, transcripts, safety identifiers, session-negotiation and network data, and model output for the public AI previews.
- Cloudflare proxies, delivers, protects, and observes public HTTPS traffic. Durbash-operated infrastructure runs the application and its API rate limits.
- Email, scheduling, signup, or application providers may process information when you follow an external destination configured by Durbash. Their notices govern information you provide there.
Review OpenAI's API data controls and Cloudflare's privacy policy for their current practices. These services may process data in countries other than your own.
Storage and retention
The current Durbash public-site application has no persistent conversation database. Chat and voice transcripts are held in browser memory. Resetting chat, starting a new voice session, or reloading the page clears the corresponding browser state; ending a voice session leaves its transcript visible until a new session or page reload.
Volatile server memory and origin rate-limit systems process pseudonymous identifiers to control abuse. Their lifecycle depends on the running process and configured infrastructure behavior. Infrastructure and provider logs may also exist under the applicable service configuration and provider policy. Durbash does not promise zero retention.
If you email Durbash, correspondence may be retained for as long as reasonably needed to respond, maintain business records, resolve disputes, or meet legal obligations.
Your choices and contact
- Do not start the AI previews if you do not want text or audio processed by OpenAI.
- Deny or revoke microphone permission in your browser, mute the microphone, or end the voice session at any time.
- Avoid sensitive information and use the reset or reload controls to clear the visible browser transcript.
- Depending on applicable law, you may request access, correction, or deletion of personal information held by Durbash.
Send privacy questions or requests to [email protected]. Durbash may need enough information to understand and verify the request.
The public demos are intended for business users aged 18 or older. Review the Terms of Use and Security page for additional boundaries and controls.
Changes to this policy
Durbash may update this policy as the public website, service providers, or legal requirements change. The date at the top identifies the current version. Material changes should be reflected here before or when the related processing begins.