We build multi‑agentic AI platforms
Using the latest multi-agentic frameworks like LangGraph, CrewAI, AutoGen and OpenAI Agents SDK, and multi-agent communication protocols like A2A, we build enterprise scale multi-agentic platforms.
Explore Kognie, our multi-agentic graphic design system ›About us
Research that became a product
Docunexus was founded in 2008 to work on document analysis. Today our main application is Kognie, a multi-agentic graphic design system, based on our founder's research on visual language models and multi-agentic systems.
Kognie
Name the goal and the platform — social, email or web — and refine the result in conversation. Every result stays editable, so you can refine it in place instead of starting over.
Who it's for
Good design for non-designers
Kognie is built for people who need visual work done well and for designers who want to move faster.
Marketers
Campaign assets across social, email and web without waiting on a design queue.
Small businesses
Announcements, promotions and brand materials, made in-house whenever you need them.
Content creators
Platform-shaped visuals that stay consistent across everything you publish.
Agencies
Faster first drafts on client work, so designers have more time for the creative calls.
Our approach
A design team, not a single model
Each agent handles one part of the process, and you steer them all in plain language.
Layout
Interprets your goal and the platform it's for, then arranges the design around it.
Style
Handles type, color and treatment so the result holds together visually.
Critique
Reviews each draft before it reaches you.
Conversation
Ask for changes in plain language. No design terminology required.
Leadership
Shekhar Pradhan
Founder and CEO
Shekhar holds a Ph.D. in computer science from the University of Maryland, College Park, and works on large language models, natural language processing and vision language models.
He has held faculty positions at several universities, including Brown University and Columbia University. At Brown he founded and led the Conversational AI Lab, where he researched foundation models and vision language models.
Contact