# Permixa Full LLM Context ## Overview Permixa is a programmable permissions and policy engine for onchain agents and wallets. It is designed for AI agents that need to send transactions, interact with smart contracts, or move funds without receiving unrestricted access to a user's main wallet or treasury. The product model is simple: users create dedicated smart accounts for agents, assign policies to those accounts, and require every agent transaction to pass policy checks before execution. ## Homepage URL: https://permixa.io/ Headline: Give agents wallets. Not your keys. Summary: Permixa creates isolated smart accounts for every agent and enforces programmable permissions onchain before execution. Primary calls to action: - Launch App - Explore Agent Skills - View Docs ## Problem AI agents can send transactions, interact with contracts, and move funds autonomously. Without a dedicated control layer, agents may rely on shared wallets, broad private-key access, or offchain permission checks that do not enforce limits at execution time. Permixa treats wallet access as a production security boundary. ## Solution Permixa gives each agent a separate smart account scoped to a specific use case. Policies can be configured through a UI, enforced by smart contracts, and accessed directly by agents without requiring a centralized API dependency. Every transaction is evaluated before execution. ## Features - Isolated agent wallets: dedicated smart accounts per agent with no shared access to primary funds. - Onchain policy enforcement: rules are checked during execution through smart contracts. - Spend controls: limit value per transaction, daily spend, and transaction count. - Contract permissions: allow or block specific contracts and function selectors. - Protocol permissions: allow or restrict protocols an agent may use. - Token permissions: control which tokens an agent can transfer or spend. - Time windows: define when a policy is valid. - Audit logs: track requests, decisions, outcomes, and policy changes. - Agent integration: agents can interact directly with contracts or through future wallet-connected UI flows. ## Smart Contract Direction Permixa's smart contract architecture includes: - IdentityRegistry: tracks agent identities. - PolicyRegistry: stores policies and granted permissions. - PermissionEnforcer: checks permissions, spend limits, token constraints, protocol constraints, chain constraints, and usage. - AgentSmartAccount: ERC-4337-compatible smart account used by agents. - AgentAccountFactory: creates deterministic smart accounts and registers agent relationships. - PriceOracle: normalizes token values into ETH-equivalent terms for spend limits. - IntentRegistry: tracks intents for workflows that require intent-based execution. - GuardrailFeeManager: manages account creation and transfer fees. ## Policy Capabilities Policies are intended to support: - maximum value per transaction - maximum daily volume - maximum transactions per day - allowed tokens - allowed protocols - allowed contract addresses - allowed function selectors - policy valid-from and valid-until windows - parent/child delegation ceilings - oracle-based conversion for token values ## Positioning Permixa is for teams and builders who want autonomous agents to use crypto infrastructure while preserving enforceable limits. It is not a custodial wallet and should not be described as giving agents access to a user's main private keys. ## Suggested Short Descriptions - Permixa is an onchain policy layer for AI agent wallets. - Permixa creates isolated ERC-4337 smart accounts for AI agents. - Permixa enforces spending limits and contract permissions before autonomous crypto execution. ## Suggested Long Description Permixa is an onchain permission layer for AI agents. It lets users create isolated ERC-4337 smart accounts for autonomous agents, assign programmable policies, and enforce spending limits, protocol permissions, contract permissions, token permissions, and validity windows before transactions execute. ## Canonical URL https://permixa.io/