Turn idle silicon into AI yield
A privacy‑first, decentralized inference network where GPU/TPU owners earn for low‑latency, accurate, high‑uptime AI serving — and developers get an OpenAI‑compatible API without data lock‑in.
SLO tier
Gold
p95 ≤ 600ms cached / ≤ 2.5s LLM
SLO tier
Silver
p95 ≤ 1.5s cached / ≤ 5s
SLO tier
Bronze
best‑effort, high availability
Privacy‑first by default
No training on user data. BYOK encryption, tenant‑segmented logs off by default, private peering option.
OpenAI‑compatible API
Drop‑in endpoints for chat, completions, images, video, embeddings, audio, and functions.
Per‑SLO pricing
Pick Gold, Silver, or Bronze for latency, accuracy, uptime, and throughput. Pay for what you need.
Verifiable performance
Proof‑of‑Inference with canaries, replica spot‑checks, Merkle‑logged traces, and attestations.
Regional pinning
Select regions like Frankfurt, Singapore, São Paulo and keep data within residency boundaries.
Hardware rewards
TEE‑attested nodes earn +5%. edged.ai TPUs based on our IP cores earn +10% reward multipliers.
OpenAI‑compatible endpoints
/v1/chat/completions/v1/completions/v1/images/v1/videos/v1/embeddings/v1/audio/v1/functions
Quick start
curl https://api.edged.ai/v1/chat/completions -H "Authorization: Bearer $EDGE_API_KEY" -H "Content-Type: application/json" -d '{
"model": "gpt-oss-20b",
"messages": [{"role":"user","content":"Write a Rust function to parse JSON"}],
"tier": "gold",
"region": "frankfurt"
}'Fee split
80%
Provider PayoutFee split
10%
Base BurnFee split
10%
Treasury/SubsidyRun a node. Earn EDGE.
One‑click installer. Auto‑model sync. QoS scoring. Non‑custodial payouts. TEEs boost rewards; edged.ai TPUs earn +10% multiplier.