Skip to main content
Sage is an autonomous second signer on your Squads Protocol V4 multisig vault. It learns how your vault transacts — amounts, recipients, tokens, timing — and screens every pending proposal against that behavioral profile, your policy rules, and external security data before anything lands on-chain. Safe transactions auto-execute. Borderline ones are surfaced over Telegram for one-tap review. Clearly malicious ones are blocked.

Sage App

Sign in, see your vault portfolio, and send gasless transfers screened on every action.

Technology

Deep dive into the architecture — Squads V4, sponsored transactions, the risk engine, and the Nanobot agent.

Quickstart

Run Sage locally — client, server, and agent — in minutes.

API Reference

Integrate with the Sage server API for sponsoring, queuing, screening, and execution.

​
Why Sage?

Multisigs solve key management, but they don’t tell you whether a proposal is safe to sign. A co-signer who blindly approves is just a slower single signer. Once a malicious proposal reaches quorum, the funds are gone.
  • A drained recipient, a fat-fingered amount, or a freshly-minted honeypot token all look identical to a wallet
  • n-of-m signers approve based on trust, not analysis — there is no screening layer between propose and execute
  • Agentic wallets need a co-signer that can reason about risk, not just hold a key
Sage is that co-signer. It holds a member key on your Squads vault with Execute permission and refuses to push anything through that fails its checks.

​
How It Works

Sage adds an AI co-signer to your Squads vault. Every proposal is scored 0–100 against the vault’s learned behavior, your policy rules, and live security feeds before it executes. Thresholds are configurable per vault.

​
Built On

Sage is fee-sponsored end to end — the server pays rent and fees, so the user never needs SOL to create a vault or send a transfer.