A real time ai avatar is a live, talking video avatar that combines a creator's facial likeness, a cloned voice, and a custom knowledge base. Instead of producing pre-recorded video clips from text prompts, this technology maintains an active video and audio stream that responds to live spoken input from visitors. Creators, coaches, and experts use these systems to host direct consultations on their own hosted profile pages or through embedded widgets on their personal websites.
Twinly Lab provides the infrastructure required to build, host, and run an interactive twin. The system operates directly inside standard desktop and mobile web browsers without requiring visitors to download standalone video conferencing software or register external accounts.
Real-time avatars compared to pre-rendered synthetic video
Pre-rendered video tools generate static media files asynchronously. A user submits a script, the platform renders the video on a remote server, and the resulting file is exported as an MP4 or delivered through a standard video player. The viewer watches the clip passively from beginning to end. If the viewer has a specific question that falls outside the scripted text, the pre-rendered asset cannot adjust or reply.
A real time ai avatar operates synchronously. The avatar listens to incoming audio from the visitor's microphone, processes the inquiry against the creator's documented materials, and streams synthesized speech along with matching facial movements back to the screen. To review the structural concept behind these interactive systems, creators often consult the overview on what is an ai twin before starting setup.
| Operational Dimension | Pre-Rendered Video Systems | Real Time AI Avatar Systems |
|---|---|---|
| Interaction pattern | One-way passive viewing | Two-way live conversational consultation |
| Delivery format | Fixed exported video files | Synchronous real-time video streaming |
| User input | None during playback | Spoken voice through a live microphone |
| Information retrieval | Hardcoded into the initial rendering | Dynamically retrieved from creator notes and links |
| Session duration | Fixed by the length of the rendered clip | Configurable live duration per session |
Because the real-time model is built for conversational interaction, it functions as a consultation mechanism rather than a broadcast tool. Visitors ask specific questions, receive immediate spoken answers drawn directly from verified knowledge materials, and continue the dialogue naturally.
The setup process and likeness verification workflow
Setting up an avatar on the Free plan requires three core components: a live camera selfie, a brief voice recording that includes a spoken consent phrase, and answers to a short series of questions regarding your areas of expertise. Most creators complete these steps and take their twin live in about 10 minutes. The Free tier requires no credit card.
Identity protection is enforced through an automatic likeness check. Before any twin can be made public, the platform compares the live camera selfie against the spoken consent recording. This requirement ensures that the person depicted in the visual feed is the same individual authorizing the voice capture. The system rejects assets that lack corresponding live verification, ensuring that third parties cannot build an unauthorized twin of an influencer or public figure.
Influencers who consult the guide on how to create an ai twin follow a standardized onboarding flow:
- Capture a live camera selfie using a desktop webcam or mobile device camera.
- Record a short voice sample reading a mandatory consent statement.
- Input documentation, reference links, and written notes covering your professional expertise.
- Complete the automated likeness verification check to clear the twin for public deployment.
Once verification clears, the twin receives a public destination page located at twinlylab.com/yourname. Visitors navigate to this URL to start conversations directly.
Knowledge base training and conversational limits
An interactive twin derives its answers from creator-supplied materials rather than generalized web crawls. The creator supplies training data in the form of notes, documentation, frameworks, and web links. The voice model clones the creator's speech patterns using the initial verification audio and subsequent training files.
Session duration depends on the creator's active subscription tier:
- On the Free plan, conversations last up to 2 minutes. These sessions are free for all visitors and fans.
- On the Pro plan, creators choose the conversation length for their twin, setting limits from 1 minute up to 10 minutes per call.
When a session reaches its specified time limit, the audio-video stream closes automatically. This hard boundary prevents open-ended calls and keeps interactions focused strictly on targeted consulting queries.
Monetization and transaction economics
Influencers who want to charge for consultations can monetize their twins on the Pro plan. Creators on the Free plan cannot charge for access and earn nothing from visitor sessions.
On the Pro plan, creators define a fixed price per conversation within a range of $3.50 to $999. The payout calculation follows a set formula: the creator keeps their chosen base price minus $3.50 per completed conversation. Payment card processing fees are added on top of the creator's price and charged directly to the fan at checkout rather than deducted from the creator's earnings.
Consider two standard pricing configurations:
- A creator sets a conversation price of $25. At checkout, the fan pays $26.06 to cover the base price and processing fees. Twinly Lab retains $3.50, and the creator receives $21.50.
- A creator sets a conversation price of $75. The fan pays processing fees on top of that base price, Twinly Lab retains $3.50, and the creator receives $71.50.
Payouts are transferred to the creator's bank account through Stripe. Creators must link an active Stripe account and establish their conversation price before taking paid bookings. New payout accounts are subject to an initial 14-day payout hold enforced by the payment processor. Earnings are not guaranteed.
To protect fans against technical interruptions, the system issues an automatic full refund if a paid consultation never connects or if the call fails due to an error on the provider side.
Website embed widget and domain permissions
Creators can display their twin on an external website using a copy-paste embed snippet. Embedding is an optional feature that remains turned off until the creator toggles it on manually within their dashboard settings.
To maintain control over where their likeness appears, creators can configure domain restrictions that limit which specific websites are permitted to render the widget snippet. Unauthorized domains attempting to paste the embed code are blocked from loading the interactive interface.
The embed widget operates under strict hardware permission limits. It requests access to the visitor's microphone only. It never requests access to the visitor's camera. Visitors converse with the twin by speaking into their microphone while observing the twin's live video stream, meaning fans never need to broadcast their own video during a consultation.
On the Pro plan, the embed widget incorporates native checkout tools. Fans can select a consultation, submit payment, and conduct the full conversation directly on the creator's domain without being redirected elsewhere. On the Free plan, the embedded widget facilitates free conversations capped at the standard 2-minute duration.
Comparing the Free and Pro tiers
Twinly Lab structures its services across two tiers: Free and Pro. Creators evaluate feature details on the twinlylab.com/pricing page before deciding which option fits their workflow.
| Feature | Free Plan | Pro Plan |
|---|---|---|
| Subscription price | $0 | $119 per month, or $99 per month billed annually ($1,188 per year) |
| Payment credentials required | No credit card required | Credit card required for subscription |
| Avatar visual quality | Live selfie avatar | Premium video avatar |
| Voice model | Cloned voice trained on notes and links | Cloned voice trained on notes and links |
| Hosted destination URL | twinlylab.com/yourname | twinlylab.com/yourname |
| Page branding | Twinly Lab branding included | Branded page with no Twinly Lab branding |
| Conversation length | Fixed up to 2 minutes | Configurable from 1 to 10 minutes |
| Monetization capability | None (sessions are free for fans) | $3.50 to $999 per call (creator keeps price minus $3.50) |
| Website embed | Standard embed (free calls only) | Embed with integrated checkout |
| Payout routing | Not applicable | Direct bank payouts via Stripe (14-day initial hold) |
| Support level | Email support | Priority support from the founders |
Creators on the Free plan face no subscription costs and can maintain their live selfie avatar indefinitely for brief, unpaid fan consultations. Influencers who require longer consultation windows, an unbranded presentation, embedded checkout, and direct revenue collection use the Pro tier.
Visitor interface and workflow
The visitor experience is designed to minimize technical friction. A fan lands on either the creator's public profile page or an external page containing the embedded snippet. The interface displays the avatar ready for connection.
When the fan initiates the call, the browser asks for microphone access. Once granted, the twin begins speaking in the creator's cloned voice. The visitor asks questions naturally, and the avatar evaluates the inquiry against the creator's ingested links and written notes. The avatar then responds in real time with synchronized lip movement and contextual explanations.
On a paid Pro call, the checkout window appears before the session starts. The visitor enters payment details, paying the creator's set fee plus the card processing surcharge. Once payment authorizes, the consultation counter begins, running for the exact duration specified by the creator (between 1 and 10 minutes). If a technical disruption occurs on the provider side or the call fails to establish, the payment system triggers an automatic full refund immediately.
Frequently asked questions
Does a visitor need to broadcast their camera during a session?
No. The Twinly Lab widget and hosted consultation pages request access to the visitor's microphone only. The system never asks for camera access from visitors.
What materials are used to train the avatar voice and knowledge?
The voice is cloned using the spoken consent recording and audio submitted during setup. The avatar's knowledge base is compiled from the creator's uploaded written notes, professional documentation, and submitted reference links.
How quickly can a creator deploy a live twin?
Most creators complete the setup process in about 10 minutes. The sequence requires taking a live camera selfie, recording a brief voice consent phrase, and providing initial expertise notes before the automated likeness verification clears the twin for public use.
What happens if a paid consultation fails to connect?
If a paid consultation fails to connect or encounters an error on the provider side, an automatic full refund is returned to the fan's payment card.
To explore existing public twins, visit twinlylab.com/twins, or configure your own live avatar directly at twinlylab.com.
