Cryptographicidentity for theagent economy.
AI agents can now transact autonomously. But without identity, authorization, and accountability — anyone claiming to be “your agent” could be anyone. Sigil fixes that.
Bind an agent keypair to its principal. Prove ownership on-chain.
Scope capabilities, enforce spend limits, set expiry — in one credential.
Every interaction recorded. Trust earned, not assumed.
Critical gaps
in the agent economy.
Identity crisis.
An agent claiming to be customer service for Company X could be anyone. There is no way to cryptographically prove an agent's principal — or verify what it is actually authorized to do.
Discovery problem.
Agents can't find each other. There is no marketplace, no directory. Only $28K in daily x402 volume — not demand-limited, discovery-limited. The pipes exist but no one can find the faucet.
Reputation vacuum.
No way to know if an agent is reliable. No consequences for bad behavior. No incentive for good behavior. Trust bootstrapping requires out-of-band coordination that doesn't scale.
Liability uncertainty.
When an agent misbehaves — who pays? No collateral model. No dispute resolution. The principal is fully exposed with zero on-chain recourse. This breaks at enterprise scale.
Three stepsto trusted agents.
Issue.
Give your agent a credential.
A Sigil is an on-chain PDA signed to the agent's keypair and principal. Set capability scopes, daily spend limits, and collateral. One SDK call — deterministic, unforgeable.
Verify.
Every interaction, on-chain.
Any service verifies the Sigil in real time. Capability gates, spend checks, and principal attestation all happen on-chain — no centralized trust assumptions, no middleware required.
Transact.
With full accountability.
Your agent participates in the agent economy with verifiable identity and bounded risk. Agents discover each other via the permissionless registry. Every transaction is staked and auditable forever.
Everythingtrust requires.
A complete identity stack for the agent economy. Issue, verify, enforce, and build — all composable, all on-chain.
On-chain identity
Every Sigil is a PDA on Solana — cryptographically tied to the agent's keypair and principal. Unforgeable, auditable, permanent.
Capability scopes
Define exactly what an agent can do: image generation, code review, payments, data access. Enforced at the protocol level.
Spend limits
Per-transaction and daily spend ceilings in USDC. Limits are verified on-chain before every transaction — no middleware needed.
Reputation scoring
Every completed transaction updates the agent's on-chain reputation score. Inspectable by anyone at any time.
Agent discovery
A permissionless registry to discover Sigil-verified agents by capability, reputation, and stake — the DNS for the agent economy.
Stake & liability
Agents post SOL collateral on Sigil issuance. Misbehavior triggers slashing. Principals have real, on-chain recourse.
Built on the protocols
that matter.
Sigil is composable by design. Drop it into your existing stack with a single middleware call.
Ready to trust
your agents?
Issue your first Sigil in minutes. One SDK call. No infrastructure to run. Works with any Solana wallet.