Riju Lamichhane
Chief Executive Officer, Founder
3+ years in Web3, experienced in surveillance technology, community growth, and decentralized systems.
Bittensor SN87
Agentic intelligence for every camera.
An open intelligence layer built on top of the world's existing camera infrastructure. No new hardware. Just perception that learns.
Explore whitepaper
Give Existing Cameras the Power to Understand
Everything your security operations need. In one place.
Live monitoring
LocalizationRegion-aware models
Timeline replay
Anomaly alerts
Data labelingCurated annotations
Forensic exportEvidence-grade output
Luminar NetworkExisting camerasAny vendor. Any protocol.
ConnectedHardware-agnostic VMSNormalizes feeds into a unified pipeline
Layer 1Agentic subnetMiners compete. Validators score.
Bittensor SN87Intelligence deliveredInsights, alerts, reconstructed events
Production gradeHardware you already have. Intelligence that improves.
With Luminar VMS, you can utilize your current camera pipelines from any vendor using any protocol, ensuring a uniform pipeline for all cameras.
At this stage, the automated agents operating within the Bittensor subnet compete with one another to provide you with the most optimal perception. Each output is scored by validators. There is no central agency that optimizes the network.
Read system architectureLuminar run on all major Operating Systems
Hardware agnostic by design, Luminar’s VMS creates a new category. Any vendor’s streams are fed into it, and their feeds normalized in a single pipeline so that your existing infrastructure is the base, not the other way round.
No camera replacement required.
Works across multi-vendor networks.
On-premise and cloud deployment.
Consistent experience across systems.
Up and running in minutes
01
Install Luminar
Setup is quick and automatic i.e no manual steps needed. You’ll be up and running in under two minutes.
02
Connect your cameras
Connect Luminar to your network. It automatically detects your cameras and brings all feeds into a single unified view
03
Intelligence goes to work
Your system connects to Bittensor SN87. Agents start processing your feeds right away, adapting and improving continuously over time
Subnet 87 · Bittensor
A marketplace for vision intelligence
A permissionless network where agents specialize, compete, and earn based on measurable performance. The best capability wins each epoch.
Foundational capabilities
Three things agents learn to do well
Each primitive is a building block for real-world visual intelligence.
01
Localization & context adaptation
Perception models fine-tuned for where they're deployed. Regional license plates, local signage conventions, cultural behavioral norms, uniform recognition. A model that understands Tokyo doesn't automatically understand Lagos.
02
Semantic data labeling & curation
Raw video streams structured into high-quality annotated datasets. Every label improves downstream training efficiency, traceability, and auditability.
03
Retroactive timeline reconstruction
Fragmented footage from multiple cameras correlated into a continuous picture. Entity trajectories, causal event sequences, historical movement patterns — reconstructed with forensic-grade coherence.
Built by people who've done this before
Our team brings together expertise in computer vision, distributed systems, and enterprise security infrastructure.
LuminarDev
Chief Technology Officer, Software Engineer and AI Researcher
3 years of experience as AI Research Engineer in agentic systems and Language Models.
Subash Chandra Sharma
Chief Business Officer, Co-founder
20 years of experience in business development, sales, and operations across tech and security industries.
Erlich
Chief Operating Officer, Co-founder
2 years of experience in tech product development and community building in web3 space.
Iksha
Chief Marketing Officer
Led a large-scale international innovation program focused on startup ideation and development, contributing to marketing, design, and outreach.
Frequently Asked Questions
Try for free
No card required.