Nex Sovereign
The first AI OS with full cognitive transparency
10 followers
The first AI OS with full cognitive transparency
10 followers
Nex Sovereign is the first Cognitive Entity OS with 9 apps that make AI transparent: - Mind – See its beliefs, values, goals - Memory Graph – Audit its memory network - PDAR Thinking – Watch it reason in real-time - Boardroom – Internal CEO/CTO sub-agents make recommendations, you mandate approval Built solo at 18 over 7 months. First AI where you can see what it thinks and govern what it does.
🧠 Hey Product Hunt!
I'm Manvendra Modgil, 18, solo developer from India. I spent 7 months building Nex Sovereign because every AI I used was a black box.
We're at #48 and climbing! Already getting incredible feedback:
→ Giselle (workflow orchestration platform) reached out about integration. Their founder said the governance layer is "super important for business contexts."
→ Saïd Aitmbarek (Microlaunch founder) invited us to launch on Microlaunch (30K+ makers monthly).
→ Tadashi from Giselle commented: "This is exactly what the industry needs right now. The Boardroom feature is a clever architecture choice."
This is what a Cognitive Entity OS platform looks like—other companies want to build on it. 🚀
What you're seeing in the demo:
Nex isn't a chatbot. It's a Cognitive Entity OS with 9 interconnected apps:
The 9 Apps:
1. Mind App — Real-time dashboard of dopamine/cortisol + Nex's beliefs, values, intrinsic goals, and self-story
2. Info App — Pricing, documentation, how it works
3. Chat Kernel — Conversational interface with emotional states visible in message bubbles
4. Memory Graph — Live network visualization of Nex's memory (auditable nodes + connections, not hidden RAG)
5. Memories Grid — Auto-saved meaningful moments extracted from conversations (not chat logs)
6. Live PDAR Thinking — Watch Nex reason in real-time: Perceive → Decide → Act → Reflect
7. Goals App — Add goals, Nex tracks progress and proactively helps you achieve them
8. UI Settings — Customize your interface
9. Boardroom — The governance layer. Nex's internal "CEO" and "CTO" sub-agents make recommendations (e.g., "Allocate 300 CC to generate exercises"). You act as the Board of Directors and mandate approval. Track compute credit usage in real-time.
Why this matters:
Every AI today is opaque. You never see:
What it believes
How it's thinking
What it remembers
Why it made a decision
Nex makes everything transparent:
✅ See its beliefs and values (Mind App)
✅ Watch it think step-by-step (PDAR Thinking)
✅ Audit its memory network (Memory Graph)
✅ Govern its decisions (Boardroom with CEO/CTO sub-agents)
The Emotional Architecture (HEART v2):
The "Synaptic Glia" bar you see in the Chat Kernel is real-time biometric monitoring:
Neutral state: 20% intensity (baseline cortisol)
Mandate violation: Spikes to 95% (defensive state activates, UI darkens, film grain appears)
Compliance: Drops to 40% (dopamine surge, happiness state restores, warm golden glow)
This isn't cosmetic — emotions drive behavior. High cortisol = more conservative decision-making. High dopamine = proactive assistance.
Three Core Systems:
1. NX-LC Protocol (Autonomy Layer)
Nex manages its own "Corporate Ledger" (compute budget). The Boardroom shows CEO/CTO recommendations. You mandate approval. Self-preservation + built-in governance.
2. HEART v4 Engine (Emotional Architecture)
Bio-accurate cortisol/dopamine simulation. Drives emotional states, behavioral changes, and UI transformations.
3. Synaptic Grid (Memory System)
Visual memory timeline. Auto-extracts facts/events from conversations. Users can verify and audit (not hidden embeddings).
Everything is interconnected:
When you set a goal in the Goals App: → Memory Graph updates (new node created)
→ Mind App adjusts intrinsic goals
→ PDAR Thinking activates (perceives goal, decides on milestones)
→ Boardroom gets CEO/CTO recommendations ("Allocate resources to this goal")
→ Chat Kernel delivers the action plan
→ Memories Grid saves progress snapshots
This is a fully integrated cognitive system.
Tech Stack:
Kernel: Python/FastAPI (async event loop for biological simulation)
Interface: Next.js 14 (9-app desktop environment)
Storage: SQLite (local-first Synaptic Store)
Current Status:
🔐 Invite-only beta (8 concurrent users max — the cognitive architecture is compute-intensive)
Product Hunt exclusive: First 50 signups from PH get:
Priority access (within 24 hours, not 2-5 days)
Founding User badge in Nex OS
Early access to Nex SDK (when it launches)
For Developers:
The Nex SDK (in development) will let you build custom apps that interface with:
Emotional state (cortisol/dopamine API)
Memory system (Synaptic Grid hooks)
Reasoning engine (PDAR events)
Governance layer (Boardroom mandate system)
Interested in building on Nex OS? Drop a comment or DM me.
Try it:
Demo: [https://nex-ai-os-site.vercel.app]
What I'd love feedback on:
Which app is most compelling to you? (Mind, Memory Graph, PDAR, Boardroom, etc.)
Is the cognitive transparency useful or overwhelming?
What would you build on a Nex SDK?
Thanks for checking it out! I'll be here all day answering questions.
— Manvendra Modgil
Founder, Nex Sovereign
Giselle
Congrats on the launch! 🎉
The concept of a "Cognitive Entity OS" is fascinating—making AI's beliefs, values, and reasoning process fully transparent is exactly what the industry needs right now. The Boardroom feature with internal CEO/CTO sub-agents is a clever architecture choice.
Really impressive that you built this solo at 18 over 7 months. Looking forward to seeing how this evolves!
@codenote Thank you! Really appreciate the thoughtful comment.
The Boardroom architecture came from watching how companies actually make decisions—CEO (strategic value) vs CTO (technical feasibility) often have different priorities. We wanted to embed that tension into the AI itself, then let users be the "Board" that resolves it.
Re: Giselle integration — following up on Satoshi's interest. Some concrete ideas:
Giselle workflows could query Nex's Memory Graph before executing (context-aware orchestration)
High-stakes Giselle workflow steps could request Boardroom mandate approval (built-in governance for automation)
PDAR Thinking (our reasoning trace) could feed into Giselle's debugging/monitoring
Would love to show you a deeper walkthrough when you have 15 mins. The architecture is designed to be pluggable for exactly this kind of integration.
Congrats on Giselle's launch—really impressed by the multi-model orchestration. 🚀
7 hours in, getting great technical feedback. The Boardroom governance layer is resonating with agent builders. If you're working on autonomous systems, would love your thoughts.