Aadhaar (UIDAI) 6
KYC & Onboarding 3
Digital Signatures 2
Data Privacy (DPDP) 2
MOSIP (National ID) 1
Products
Video KYC · AI video interviews · Video personal discussion
Verify customers, interview candidates and run credit personal discussions over video — with your own agents, an AI agent, or both, in the browser or in your mobile app.
One platform, two kinds of agent
The same customer screens, the same recordings, the same review queue — whoever is on the other side of the call.
Never a dead end. A customer set for the AI agent moves to the live-agent queue automatically if no AI slot is free — and both kinds of call end in the same checker queue.
Channels
Nothing to install for a browser call, or a native app with guided capture — on the same back end.
Virtual Agent
Please do not switch to any other application until the video call is completed
Camera, microphone and location are checked first. Location uses GPS where it can and falls back gracefully where it can't.
A live connection-quality indicator, automatic reconnection and a screen that stays awake through the call.
English, Hindi and Tamil live today, with Telugu, Marathi, Kannada and more ready to switch on — every screen, every spoken line.
Your name, logo and copyright line across the web and the mobile app, in light or dark theme.
Live agent
The customer, their documents, the checks and the script — with nothing to switch between.
03 Nov 2026 11:42:18 · 12.9716 N, 77.5946 E
Agents mark themselves ready and are matched with the customer who has waited longest.
A full-resolution face capture is matched against the customer's photo and checked for liveness during the call.
Snapshots of agent and customer with the time and location stamped on, kept with the record.
Questions appear on the customer's screen and are marked matched or not matched as the agent goes.
Background blur in three strengths, preset office images or your own, remembered for each agent.
Every call ends as complete, or not complete with a reason, and moves to review.
AI agent
A lip-synced virtual officer runs the whole conversation in the customer's language — and checks every answer as it goes.
During the call
The officer greets the customer, explains the call and captures their face photo and location.
Each question is spoken with natural lip-sync and shown on screen; the answer is transcribed as it is spoken.
Names, amounts, dates and counts are read from the answer — and a follow-up is asked if a detail is missing.
Answers are compared with your data, and the call, each answer's audio and the transcript are kept.
A realistic officer with lip-synced speech. Replace the officer from the admin screen with a short silent video — checked, previewed and switched on in a few steps.
Rule-based checks for numbers, dates and Indian names, backed by an AI model that reads meaning — on your own servers.
Re-asks when nothing is heard, asks again slowly for numbers and dates, and follows up when a detail is left out.
Standard questions plus lines written for each case — and branches that follow earlier yes-or-no answers.
The officer can ask the customer to show their business or home on camera; the walk-through is kept as its own clip.
Set how many AI calls run at once. When all are busy, customers are told in their language to try again, and the link stays valid.
Solutions
Each runs on the same agents, channels, recording and review — configured for its job.
Verify customers over video with a live agent or the AI agent, and approve each record through maker–checker review.
The credit personal discussion, run by the AI officer and scored against the application data and the customer's proof documents.
Pre-joining and referral interviews conducted by the AI interviewer, graded automatically and reviewed by your team.
11:42 · GPS| Question | Said | On record | Score |
|---|---|---|---|
| Full name | Anitha Gowda | Gowda Anitha | 92 |
| Loan purpose | Expand my tailoring unit | Business expansion | 100 |
| Monthly income | 62,000 | 65,000 | 90 |
| Family members | 4 | 5 | 60 |
| House proof owner | In my name | From document | 95 |
| Premises walk-through | Video recorded | — | Recorded |
The right screen for every role
Both sides of the call merged into one video with the time and location burned in, a marker for each question, and every answer's audio on its own.
Create cases through a secure API and receive results, transcripts and recording links by signed callback, retried until delivered.
Security and deployment
Installed on your own servers, from a two-server setup to a full three-tier layout with disaster recovery.
Video, recordings, databases and the AI models run in your data centre; only speech-to-text and voice use an external speech service.
Role-based screens and endpoints, captcha at sign-in, lockout after repeated failures, forced password change and session timeouts.
Strict content security policy, HTTPS everywhere and protective security headers on every page.
API keys stored only as hashes, optional IP allow-lists, rate limits and signed callbacks.
Call times, locations, agent outcomes and every checker action are kept with the record.
Add media and recording nodes as volumes grow; recordings are merged by a queue that retries and never duplicates.