CollaborativAI

Multi-AI consensus with barycenter analytics, Voronoi partitions, and round-by-round debate orchestration

We pair frontier AI engines with geometric convergence metrics, rationale sharing, and transparent debate records so research teams can interrogate ideas and watch agreement emerge in real time.

Try Maestro 3

Learn More

Our Mission

We are actively building the CollaborativAI R&D engine: Enhanced V2 features with Venn overlap analysis, barycenter proximity clouds, convergence proofs, and rationale-aware prompting. Our internal Lab (Team2025) dogfoods every build, running structured debates across GPT-4, Claude Opus, Gemini 1.5, Grok, and more before promoting modes to the public workspace.

Scientific Rigor

Geometric consensus metrics (barycenter vectors, Voronoi partitions, convergence proofs)

Transparent Methods

Structured prompts with rationale + source sharing in every debate round

Open Collaboration

Team Lab with shared API vault, Google Drive library, and discussion forum

Research Areas

Enhanced V2 Analytics

Venn overlap, proximity clouds, and convergence metrics that track semantic and evidential alignment

Barycenter & Voronoi Visuals

Geometric partitioning that reveals how AI engines claim semantic space across debate rounds

Lab Dogfooding

Passcode-protected Lab console for internal experiments, rationale reviews, and rapid iteration

Google Cloud Platform

Cloud SQL Postgres, Drive integration, and shared API vaults keep the research workspace in sync

Dynamic Mode Registry

Simple, Enhanced V2, V3 Research, and experimental V4/V5 toggles ship via a config-driven pipeline

Debate Orchestration

Transformer-inspired cross-interrogation that steers engines via barycenter vectors each round

What's Live Right Now

Enhanced V2 Workspace

Run multi-engine debates with live Venn diagrams, barycenter proximity clouds, convergence curves, and rationale/source summaries. Watch semantic and evidential distance shrink as engines align.

• Works with GPT-4, Claude Opus, Gemini 1.5 Pro, Grok, DeepSeek, Mistral
• Team defaults auto-load shared API keys (Lab passcode required)
• Source-aware prompts capture citations every round

Protected Lab & Library

Internal dogfooding hub with Google Drive library, discussion forum, and experiment console. Every experimental mode ships here first before graduating to the public app.

• Passcode Team2025 unlocks Lab during active development
• Live Drive listings with owners, modified dates, and quick links
• Forum threads tied to convergence runs and shared documents

Research Papers

PUBLIC

Multi-AI Convergence Analysis: Methods and Applications

A comprehensive framework for analyzing consensus across multiple AI systems using semantic similarity and graph-based approaches.

Paolo Pignatelli et al. • Coming Soon • arXiv preprint

TEAM ONLY

Team Research Papers & Working Drafts

Access protected research papers, working drafts, and team collaboration materials.

Team Login Required

Our Team

Res Saleh

CEO

Corporate Development

Paolo Pignatelli

CTO / EVP R&D

AI Consensus Methods, Multi-Model Analysis

Jody Fast

VP Marketing & Sales

Business Development & Strategy

Lynn Saleh

Special Advisor to CEO

Strategic Planning & Advisory

Team Workspace

Collaborate on research, share materials, and discuss ideas in our protected team space. Future: AI agents will participate alongside team members!

Access Team Space

Get In Touch

Interested in our research or potential collaborations?

Paolo@Collaborativ.ai