1,000,000,000 $TEMNO total supply. 20 halving epochs over ~20 years.
Emission (Node Rewards)60%
20 halvings, starts 30K/hr
Public Sale10%
100M tokens at $0.009
Airdrop10%
Community + slash recycling
Team10%
1yr cliff + 2yr linear vesting
Marketing5%
KOLs, bounties, listings
Reserve5%
Emergency / strategic
Total Supply
1B
Decimals
9
Chain
BSC
Standard
BEP-20
How Temno Works
01
Deploy a Node
Spin up a $5-80/mo VPS. Install the Temno Daemon + Ollama. Stake $TEMNO to join the network.
02
Execute AI Tasks
Clients send inference requests. Your node runs them locally using LLMs. Earn $TEMNO per task.
03
Train Together
Nodes collect execution traces, train LoRA adapters locally, share only weight diffs. Zero data leaves your node.
04
Prove & Earn
Proof-of-Learning verifies real model improvement. Rewards scale with contribution, not just uptime.
Key Innovations
Federated LoRA TrainingPrivacy-First
Nodes train adapters locally on execution traces. Only weight differentials travel over P2P. Raw data never leaves the node. Aggregator merges via weighted FedAvg.
Model ShardingScale
4 Lite nodes (8GB each) combine via pipeline parallelism to serve 70B+ parameter models. No single node could do it alone. Coordinated via P2P cluster protocol.
Proof-of-LearningVerified
20-question benchmark challenge per LoRA submission. Before/after scores verified by aggregator. Improvement delta recorded on-chain. Prevents poisoning attacks.
Speculative RoutingSmart
Tasks classified into 7 semantic domains. Nodes build per-category specialization scores. Router picks the best node per task type, not random broadcast.