Security monitoring, investigation, and controlled response for tokenized financial infrastructure — with evidence your auditors can verify.
Capabilities
From the first signal to the evidence package, Guard keeps detection, investigation, decisions and proof in one place.
Watches the contracts, wallets and roles in your monitoring scope on supported EVM networks.
Role grants, admin changes, upgrades and other privileged actions are detected and raised for review.
Severity-based routing and escalation; related alerts are correlated into one incident with a timeline.
Evidence-grounded summaries and runbook suggestions. Recommend-only: AI never takes an action itself.
A rule-based policy engine evaluates every recommended response. The same input gets the same decision.
Approval quorums and step-up MFA. Production execution is off by default on every plan.
Evidence packages signed with Ed25519 over a SHA-256 manifest, verifiable offline. Audit logs are hash-chained.
Workspaces, role-based access, a shared Decoda identity with MFA, and workspace-scoped data.
How Guard works
AI recommends, a deterministic policy engine decides what needs approval, and people authorize. Every step leaves a record.
Typed detections for privileged activity, suspicious transfers and policy violations.
Incidents with forensic timelines and AI-assisted, recommend-only summaries.
Deterministic policy decides what needs approval; the right people approve.
Your team acts through its own processes and signers. Nothing runs automatically.
Signed evidence packages and a hash-chained audit trail of every decision.
The pilot boundary
Pilots keep the full security workflow and lose exactly one capability: production execution. The boundary is enforced server-side, before any provider call, and every refusal is written to the audit log.
Security architecture
Guard’s JSON-RPC access is read-only: it never calls a signing or transaction-sending method.
Decoda does not hold customer wallet private keys or seed phrases, and refuses submissions that look like them.
Monitoring, telemetry, alerts, incidents and evidence are scoped to a workspace end to end.
Missing or degraded data is shown as missing or degraded. No alert is never presented as proof of safety.
Plans
Pilot: a scoped evaluation, by request. Scale: from $999 per month. Enterprise: custom pricing.
Scoped evaluation
Tailored
Contact us to scope your evaluation
Validate Decoda RWA Guard against your own infrastructure with a defined scope and guided onboarding.
Ongoing security operations
From $999/ month
Final pricing depends on monitoring scope
For teams moving beyond pilot evaluation to recurring security operations.
Institutional programs
Custom
Custom pricing
For institutions with a broader monitoring scope and their own security and governance requirements.
FAQ
A defined monitoring scope — one workspace and up to five monitored contracts — with guided onboarding, alerts, incident investigation and evidence workflows. Pilots are recommend-only: Guard cannot execute production changes during a pilot. Scope and terms are agreed with your team before the evaluation starts.
On every plan, Decoda recommends and your people authorize. Production execution is off by default and is enabled only for an Enterprise tenant through an explicit, audited agreement. Decoda never holds customer private keys.
Guard monitors supported EVM networks. Tell us which networks and contracts you need covered and we will confirm before your pilot is scoped.
AI-assisted triage summarizes incidents, builds evidence-linked timelines and suggests predefined runbooks. It has no tools that act on-chain, and every recommendation needs human approval. Policy decisions are made by a deterministic engine, not by a model.
Not today, and we do not claim them. We are happy to walk your security team through Guard’s architecture, controls and evidence model — contact us.
Request a scoped pilot. The Decoda team reviews every request and agrees the monitoring scope with you before anything is provisioned.