Meetra AI - API for Human Interaction Analytics - Sell Project

Meetra.ai materialized within the growing AI landscape to provide human conversation insights based not only on context but also HOW humans interact Its mission is to help understand human groups in the conversational settings. The ultimate goal is to analyze meeting audio and improve overall group and service efficiency through valuable insights otherwise unavailable with just plain context, or transcript of the conversation. Project Structure: Independent Project Product: AI API for human conversation analysis Team : Founders (Andrzej Dobrucki, Mikolaj Skubina), Matt Kozowski (UMCS), Developers at UMCS Industry: Artificial Intelligence: Advanced voice analysis and processing Financial Sourcing: Venture Capital and NCBR (National Center for Research and Development) 1M PLN Grant in collaboration with UMCS University Poland and Politechnika Lubelska Development status: API Platform up and running (http://api-test.meetra.ai) TRL 9 Technical Readiness Level Ready to use and build upon Deployable anywhere in under 1hr Scientific Paper in writing by Politechnika Lubelska Reason for Sale: Strategic refocus - the company is focusing on its proven business model of hardware manufacturing & distribution with its limited human resources Key Highlights USP: Own cloud or on-premise PaaS audio-processing AI platform to generate conversation insights and delve into sophisticated human interaction dynamics. Example Use Cases: Everywhere where insights can be obtained from an audio material of a human conversation. Examples include: Team Meeting analysis for remote team efficiency improvement Sales Call analysis for sales strategy efficiency Customer Support Calls analysis for customer experience improvement Political Debate analysis to understand speaker dynamics and trends Doctor - Patient interview analysis for customer experience quality improvement Pitch and VC interview analysis to save time and money picking the best potential startup Technology/IP: Meeting & group dynamics analytics with modules including: State-of-Art features Speaker recognition Conversational sentiment (Positive, Negative) Conversation balance - talk : listen ratio Conversation energy over time Weighted interaction graphs between speakers High quality bilingual transcripts (English + Polish), any language can be added Auto-recognition of conversation topics and per topic characteristics Topic energy Topic involvement Topic sentiment Topic participants Speaker and conversation emotion mapping Emotion recognition (speech tone and context based) Emotion timeline over time Question detection Speaker audio quality detection (3-point scale) Architecture  Multi-threaded ensemble model operating simultaneously 19 independent API endpoints for retrieving analytical data from processed audio material Full data security with 2048 bit database encryption On-premise or in-cloud placement, all modules are API-accessible Paired with open LLMs such as Llama3 provides human-like interpreter for raw conversation insight data Support for fast GPU Processing Backend: Python3, Flask AI : PyTorch, StableDiffusion, Huggingface Code Quality University grade code quality easy to extend and scale Fast deployment in under 60 minutes anywhere Complete documentation Software prototype (Meetra BOT) Linux-based Meetra BOT user can invite to a Google Meeting to record and perform further analysis of recorded material on the Meetra Server Meetra server for further analysis. Screen below Additional notes Detailed information and API access - https://linktr.ee/meetraai Makers available to answer questions and support deployment, implementation All R&D results were audited and its quality rigorously tested and confirmed with algorithm reliability > 90% according to NCBR requirements Models can be easily trained using own contextualised data