tools
Tavus Conversational Video Interface
Tavus Conversational Video Interface lets developers embed real-time AI video conversations with voice, vision, memory, and digital personas.
In inglese
Tavus CVI is an API and developer platform for building real-time, face-to-face AI conversations. It combines speech, language, perception, rendering, orchestration, and WebRTC, with support for custom personas, knowledge bases, guardrails, memory, and tool use.
Teams use it for sales assistants, interviewers, healthcare intake, learning, coaching, support, and product experiences. A free plan is available; paid plans include monthly fees and usage-based overages. Enterprise features such as custom concurrency, white-labeling, and SLAs cost extra.
Features
- Embed real-time conversational video and voice through APIs
- Use AI personas with custom faces and voices
- Process voice, video, and screen context
- Add knowledge bases using retrieval-augmented generation
- Configure memory, objectives, guardrails, and function calling
- Review conversation transcripts, recordings, and perception analysis
- Support audio-only and video conversation modes
- Deploy with up to 10 concurrent streams on Growth
Use cases
- Build sales assistants that answer product questions on websites or calls
- Screen job candidates with conversational AI interviewers
- Provide patient intake, education, and support
- Create tutors, coaches, and employee training agents
- Guide users through products with live screen and browser interaction
- Embed branded AI video agents in customer applications
Pros
Cons
Latest updates
- 🧪 Test mode for Conversations
You can start conversations in test mode, where the replica does not join.
- 🔍 Fuzzy search
Fuzzy search finds personas based on partial matches of UUIDs or names.
- Teaching machines how to be human
Memories, Knowledge Base, Objectives & Guardrails, and Persona Builder are live in the portal.
- 🧩 Events console now live in dev portal
An events console makes it easier to monitor everything happening during a conversation in real time.
- 🦄 Conversation transcripts + perception analysis
You can view conversation details in the developer portal, including transcripts with speaker roles and perception analysis.
Capabilities
Get it
Pricing
- Starting price
- $20/mo
- Prices checked
- 2026-09-26
Free
Free
- Unlimited messaging
- 15 minutes of voice & video calls
- Support for 30+ languages
Plus
- $20/mo
- Unlimited messaging
- 150 minutes of voice & video calls
- Support for 30+ languages
- MCP Early Access
Max
- $50/mo
- Unlimited messaging
- 500 minutes of voice & video calls
- Support for 30+ languages
- MCP Early Access
Basic
Free
- 25 minutes of AI conversation — full CVI (video) or voice-only, included
- Access to 25 stock AI humans
- Whitelabeled API access
- Support for 30+ languages
- Developer docs + community
Starter
- $59/mo
- 100 minutes of AI conversation (CVI)
- 3 custom AI humans
- Up to 3 concurrent conversations
- Full visibility into conversation transcripts
- Pay-as-you-go with no overage cap
Growth
- $397/mo
- 1250 minutes of AI conversational video
- 7 custom AI humans
- Access to 100+ stock replicas
- Up to 10 concurrent streams
- Conversation recordings with full transcripts
- Advanced persona controls — memory, guardrails, custom objectives
Enterprise
Price on request
- Unlimited custom replicas
- 100% white-labeled experience
- Custom concurrency limits
- Guaranteed SLAs — instant boot times and uptime
- Enterprise-grade security and compliance
- Early access to new models and features before public launch