Live product surfaceImplemented
The website explains Looma as a verified road-alert network and includes a working operator-style demo console.
Internal technical brief
How Looma becomes an AI-operated road-alert network.
Looma is a software network for verified road alerts. It gives connected vehicles, fleets, cities, and roadside systems a trusted way to send safety messages to nearby operators, then turns the live signal trail into pilot evidence investors can inspect.
Product status
What exists, what is prototype, and what must still be funded.
Cryptographic identityImplemented
The browser creates an Ed25519 identity, signs safety messages, and sends structured payloads to the relay.
Edge relayImplemented
A Cloudflare Worker relay receives V2V messages, persists recent events, and serves nearby alerts for the console feed.
Nearby event feedImplemented
The demo can show nearby road alerts using GPS when available and a deterministic demo location when GPS is blocked.
AI and identity servicesPrototype
The backend includes FastAPI services and ML-oriented endpoints that can become the intelligence layer around alert quality, evidence, and reporting.
Certified deployment pathNot certified yet
Looma should not be represented as a certified road-safety, C-V2X, or SCMS production system until those integrations and approvals exist.
Network design
The core architecture is a trust pipeline, not a dashboard skin.
Signals
Vehicles, fleet devices, phones, or roadside units emit road events such as hazard, slowdown, obstruction, or emergency proximity.
Trust
Looma attaches identity, signatures, replay protection, and origin metadata so the network can separate real events from anonymous noise.
Relay
The edge relay accepts signed alerts, keeps a short operational window, and returns location-relevant events to nearby operators.
Agents
AI agents summarize clusters, identify suspicious patterns, draft operator briefs, and prepare pilot evidence without controlling vehicle motion.
Console
The operator console turns raw alerts into an intelligible control-room view for pilots, integrators, city teams, and fleet managers.
FinsavvyAI ecosystem
Where AI agents, generative AI, MCP, and the partner platforms fit.
The credible story is not that Looma magically becomes autonomous. The stronger story is that Looma exposes a safety-message network to controlled tools, wraps those tools in governance, and uses generative AI to produce evidence, summaries, risk checks, and integration work.
LunaOS.ai
Target runtime for AI-native apps: auth, APIs, storage, deployment, backend agents, and product operations.
Looma can use it as the managed operating layer for accounts, pilot tenants, agent state, and internal workflows.
OpenSyber.cloud
Governance layer for AI-assisted work: MCP control, agent permissions, contractor isolation, monitoring, and audit.
Looma can use it to approve, log, and constrain every agent action that touches safety messages or customer evidence.
ClawPipe.ai
AI pipeline layer for model routing, semantic caching, budget control, tracing, and provider failover.
Looma can use it to route generative summaries and agent prompts through the cheapest reliable model path with traceability.
Amliq.finance
Compliance intelligence for screening, sanctions, PEP, adverse media, enforcement, cases, and audit reports.
Looma can use it to screen fleet customers, resellers, municipalities, vendors, and high-risk procurement counterparties.
LunaForge
Code context and change-risk intelligence with MCP export, dependency mapping, and architecture policy evaluation.
Looma can use it to prove engineering diligence when the alert network changes before pilots or investor review.
Investor-safe framing
Do not sell the MVP as a certified road-safety system.
- Generative AI does not decide whether a vehicle brakes, swerves, or changes route.
- The MVP demonstrates signed road alerts and operator evidence, not certified production C-V2X hardware.
- Partner platforms are positioned as integration targets until wired into production workflows.
- Every investor claim should separate live demo capability, prototype backend capability, and funded roadmap capability.