Vigil
Calendar-aware planning, alarms, and real-world schedule context for daily agent workflows.
- Outlook & Teams calendar integration
- Auto-sets and cancels wake-up alarms
- Detects early meetings before you miss them
- Sleep nudges based on schedule
DailyClaw gives autonomous agents a verifiable identity, enforced policy, and local-first privacy — before they act in production. Built on ERC-8004. Zero data egress.
Reads your schedule. Sets alarms before early meetings. Cancels them automatically. No manual input required.
ETH-signed ZK commitments. Balance stays hidden. Spend limits enforced before every transaction fires.
Sieve quietly removes personal details before AI ever sees your data: names, IDs, addresses. Restored after. Fully local.
Without identity, policy, and privacy baked in, agents leak credentials, move funds unchecked, and leave zero audit trail. Every deployment is a liability. DailyClaw fixes the foundation.
Agents that can act without a cryptographic identity are unattributable and unverifiable. Without ERC-8004, there is no on-chain proof of who owns the agent or what it is authorized to do.
Your agent reads emails, documents, and messages containing passwords, API keys, and personal data. Every token goes straight to the model unfiltered — with no redaction layer in between.
Without calendar integration, your agent has no awareness of the real world. No alarms. No schedule context. No sleep nudges. You discover the 7 AM meeting when you're already late.
A single misread instruction or agent bug can cause real financial damage. Without a policy layer, there are no spend caps, no quiet hours, no domain controls — nothing checked before an action executes.
When an agent acts, there is no cryptographic record of what it did or why. Nothing to dispute a transaction with. Nothing to verify past behavior against. No accountability.
Without a standard like ERC-8004, agents have no cryptographic identity that travels across apps. No ownership proof. Every new session starts from zero with no trust history and no continuity.
Calendar-aware planning, alarms, and real-world schedule context for daily agent workflows.
Verify that an agent is real, attributable, and portable across environments.
Constrain what agents can do before any high-risk action reaches execution.
Portable reputation built from signed local logs. Cross-app and verifiable.
Signed receipts and tamper-evident logs for every agent action. Fully auditable.
Zero-knowledge payment commitments with policy checks before funds move.
Detects and strips sensitive data before the model sees it. Restores real values in output. Fully local.
Every agent action passes through a four-layer pipeline: identity verification, policy enforcement, skill execution, and cryptographic proof. Nothing leaves the machine unverified.
ERC-8004 registration, ownership proof, portable metadata. Your agent has a verifiable on-chain identity before it ever takes an action.
Spend caps, permission checks, quiet hours, and manual approval gates — all defined by you, enforced synchronously before every agent action.
Skills run — Vigil reads your calendar, Specter authorizes payments privately, Sieve strips sensitive data — all on-device with output filtering before any response leaves.
Every action produces a signed receipt and an entry in the tamper-evident event log. Persistent trust state. Full audit trail. Nothing unverified leaves the machine.
ERC‑8004 trustless agent identity, ZK private payments, and local-first execution. The foundation for safe real-world agent deployment.
On-chain trustless agent registry. Register, delegate, verify, and revoke agent identities without any intermediary. Anchors DailyClaw agents to Ethereum.
Pedersen-style commitments hide the exact payment amount from the processor. Range proofs confirm amount ≤ policy limit without revealing the value.
Cryptographic ownership proofs using standard Ethereum message signing. Portable identity across all OpenClaw apps, verifiable by anyone.
Every skill runs on your machine. No data leaves your device. PII redaction, ZK commitments, and rules enforcement are all on-device.
Full calendar awareness for your agent. Detects early meetings, sets wake-up alarms, and clears them automatically when plans change.
Three-tier detection removes personal details before AI sees your data: names, emails, API keys, addresses. 13+ entity types, token-mapped and restored after. All local.
DailyClaw is the missing layer between agent capability and real-world execution. Early access is limited to builders and early adopters participating in the next wave of autonomous agents.
No spam. Only alpha updates and product access notes.