Your AI Digital Twin
for Meetings
Vocaris joins your Google Meet meetings, learns how you communicate, and speaks on your behalf — with your voice, your priorities, and your personality.
Meetings Drain Your Best Hours
You built your career on expertise — not on sitting in virtual rooms repeating the same updates.
8+ hours/week in meetings
The average professional spends more time in meetings than doing actual work.
Context switches kill focus
Jumping between meetings fragments your deep work into useless 20-minute gaps.
You can't skip — they need YOU
Note-takers and recorders miss the point. Your team needs your voice, your decisions, your presence.
An AI that speaks as you — not just for you
Vocaris doesn't just take notes. It joins your Google Meet meetings, observes how you communicate across 10 real meetings, then builds a precise behavioral profile.
- Learns from real meetings, not descriptions
- Speaks with your actual vocabulary and tone
- Defends your deal-breakers under pressure
- Remembers the last 5 meetings for context
Meetings 1–10
Observer Mode
Joins silently. Listens. Learns your communication patterns, priorities, jargon, and decision-making style.
Meeting 11+
Active Mode
Speaks as you. Responds in real-time with your voice, priorities, and personality.
How Vocaris Works
From sign-in to autonomous meeting participation — a seamless pipeline that learns and evolves.
Sign In with Google
One-click Google authentication. Your account is secured with OAuth 2.0, and your refresh token lets the bot join meetings as you — bypassing waiting rooms.
Voice Onboarding
Have a real-time voice conversation with our AI agent. Tell it about your role, projects, communication style, and priorities. Create separate profiles for each job.
Observer Mode
For the first 10 meetings per job, the bot joins silently. It listens, stores transcripts, and uses DSPy to extract behavioral insights — learning how you actually communicate.
Active Mode
From meeting 11 onward, your AI twin takes over. It speaks with your vocabulary, enforces your deal-breakers, remembers past discussions, and gets smarter with every meeting.
Built for Intelligence
Every feature is engineered to make your AI agent smarter, faster, and more authentically you.
DSPy Behavioral Learning
A structured AI pipeline extracts communication style, priorities, jargon, and deal-breakers from your real meeting transcripts.
Voice Onboarding
Talk to a real-time voice AI agent about your role, projects, and work style with sub-second latency.
Multi-Job Support
Freelancer and full-time engineer? Each job gets its own onboarding, observer phase, and adapted persona.
Calendar Auto-Join
Connect Google Calendar once. Vocaris scans for meetings every 2 minutes and auto-joins them automatically.
Scrum Standup Mode
Activate scrum mode and the bot categorizes every update as Done, In-Progress, or Blocked in real-time.
RAG Meeting Memory
Every transcript is embedded and stored. Query any past meeting with natural language.
ClickUp Integration
Push meeting summaries and scrum outputs directly to your ClickUp workspace as trackable tickets.
Smart Auto-Leave
Three-layer safeguard ensures the bot only leaves when the room is genuinely empty for 5 minutes.
Why Vocaris
The only meeting agent that doesn't just listen — it becomes you.
Not a Note-Taker — A Digital Twin
Identity-AccurateOther meeting bots record and summarize. Vocaris actually speaks as you. It uses your vocabulary, enforces your priorities, and defends your constraints — even under pressure from other participants.
Continuous Behavioral Refinement
Always ImprovingThe agent doesn't plateau after setup. After meeting 10, the DSPy pipeline re-synthesizes your behavioral profile after every subsequent meeting. The more it observes, the more precisely it becomes you.
Learns from Real Meetings
Observed BehaviorYour persona isn't built from what you tell us about yourself — it's extracted from how you actually behave in real meetings. Logic priorities, approval thresholds, deal-breakers — all observed, not self-reported.
Per-Job Intelligence
Multi-PersonaYour tone as a lead engineer at Google is different from your freelance consulting voice. Each job has its own independent learning pipeline, observer insights, persona, and meeting history.
Powered by Cutting-Edge AI
A carefully architected stack combining the best AI services for voice, language, and memory.
AI & LLMs
Groq / LLaMA 3.3-70B
Primary LLM
DSPy + LiteLLM
Behavioral Pipeline
OpenAI Embeddings
Semantic Search
Voice & Audio
Deepgram
Speech-to-Text
ElevenLabs
Text-to-Speech
Recall.ai
Meeting Bot Infrastructure
Backend & Data
FastAPI
Async Python API
Supabase
PostgreSQL Database
Pinecone
Vector Database (RAG)
Auth & Integration
Google OAuth 2.0
Authentication
Google Calendar API
Auto-Join
ClickUp API
Task Management