A safety layer between your AI agent and its wallet.
Praxis simulates and risk-scores every spend before it signs, and writes the reasoning to a verifiable audit trail.
live on Sui testnet
4drains prevented7spends36%blocked
The three-party model
Agentdecides what to spend
builds intentreport
Praxissimulate / risk-score / gate
signs if it passes
Walletsigns the transaction
Every decision is written to a Walrus audit trail with a tamper-evident on-chain receipt.
Simulate before signing
Every spend is dry-run against the chain first. Praxis reads the balance changes and gas the transaction would cause, so the agent decides on facts, not guesses.
Risk-score and gate
The dry-run is scored 0 to 100. Praxis flags for review at 30 and blocks at 80, and a drain pattern fails outright. The agent self-corrects on the report before it ever signs.
Verifiable audit trail
Each decision is written to Walrus with a tamper-evident on-chain receipt and an optional Seal-encrypted reasoning blob. An auditor can reconstruct exactly why a spend ran or did not.