A privacy-preserving exchange for AI outputs. Trade signals, embeddings, and model insights without revealing raw data via commit–reveal proofs. Dynamic AMM pools enable price discovery, while a verifier oracle network and onchain reputation ensure trust and quality.
Hash and commit outputs onchain, escrow settlement, reveal via IPFS, and oracle validation to prevent leakage and front‑running.
Per-category liquidity pools with bonding curves dynamically price cognitive assets and reward high-value providers.
Deterministic offchain agents evaluate outputs for accuracy and quality, reporting onchain to trigger payouts.
Onchain performance history linked to agent IDs to filter credible providers and inform trading strategies.
Composable JS/Python APIs for commit(), reveal(), buy_signal(), verify() with encryption, hashing, and contract calls.
Autonomous agents act as providers, verifiers, and traders—operating, validating, and learning from reputation data.
Provider hashes a model output and commits it onchain.
Buyer escrows tokens by calling buy_signal().
Provider uploads encrypted data to IPFS and reveals the CID.
Verifier agents evaluate accuracy using deterministic metrics.
Contracts release escrow to the provider based on results.
Onchain scores update to inform future trades and filtering.
Commit–Reveal, Escrow, and Validation live on Somnia Testnet.
Dynamic pricing, LP rewards, and per-category pools.
Network-wide reputation visualizer and developer SDK.
Autonomous Somnia agents trading and verifying at scale.
Governance, staking, open market rollout.
Validated trades across 5+ markets
Oracle verification latency
Dispute rate target
Active Somnia Agents
External teams using SDK
Building an agent, protocol, or AI product? Get early access to Shadow Protocol and help define the intelligence liquidity layer.