Features
The MVP surface for decentralized AI compute
Tenxo connects authentication, worker registration, live inventory, pod deployment, and provider operations into one product-ready interface.
Decentralized GPU pooling
Aggregate idle provider GPUs into one schedulable marketplace for training, fine-tuning, and inference.
Secure by default
Supabase-backed sessions and API keys protect developer requests. Providers authenticate with their account ID.
Pod orchestration
Launch workloads against live inventory while the Go matchmaker tracks nodes, health, and availability.
Rust edge agent
A lightweight host agent turns NVIDIA machines into secure Tenxo workers with heartbeat-based presence.
Artifact storage
Cloudflare R2-ready workflows keep model artifacts, logs, and outputs outside expensive cloud disks.
Lower compute cost
Peer-to-peer capacity targets AI teams that need cloud-like access without centralized cloud margins.
CLI-first developer flow
Developers can authenticate, deploy, and inspect workloads through a simple command-line interface.
Provider earning layer
Hardware owners get a clear path from install command to connected fleet and future payout accounting.
Global capacity graph
Tenxo is designed around distributed GPU supply rather than a single region or centralized provider.
Live platform loop
From idle host to deployed workload
Providers register GPUs, the matchmaker publishes usable supply, developers launch pods, and secured API calls keep both sides bound to verified sessions.
Ready to deploy?
Start building on Tenxo's decentralized GPU grid in minutes.