Question 1

What is Daylite?

Accepted Answer

Daylite is the runtime enforcement and verifiable audit layer for AI agents in regulated enterprises. It is a self-hosted Rust binary that sits at the MCP tool-call boundary. It checks each agent action against policy before the action runs and decides whether it executes, then writes a tamper-evident, cryptographically signed record of what was requested, what policy decided, and what actually ran. It deploys in your VPC, on-prem, or in an air-gapped network.

Question 2

How is Daylite different from Vanta, Drata, Credo AI, or OneTrust?

Accepted Answer

Those tools document and score policy. They map it to frameworks and produce reports, and they sit out-of-band from the running agent, so they cannot block a non-compliant tool-call. Daylite enforces policy at the tool-call boundary at runtime and produces cryptographic proof of what the agent actually did. Keep your governance tool for documentation. Daylite sits beneath it.

Question 3

What does 'built on FIPS-validated cryptography' mean?

Accepted Answer

Daylite runs its cryptography through AWS-LC, a FIPS 140-3 validated cryptographic library. Daylite itself is not product-certified under CMVP, and we do not claim a CMVP certificate. Commercial enterprise procurement under SOC 2, HIPAA, and DORA accepts library-level validation. FedRAMP High and certain DoD contracts require product-level CMVP, which is on the roadmap.

Question 4

How does an auditor use the evidence?

Accepted Answer

Your team runs one command to export the record. The auditor verifies the hash-chain offline with the verify tool, confirms no entry was changed after it was written, and reviews the decisions. The format is open, so there is no lock-in on the evidence itself, and no trust in Daylite is required to check it.

Question 5

Where does Daylite run, and where does the data go?

Accepted Answer

Daylite is self-hosted and air-gap capable. It runs in your VPC, on-prem, or in a disconnected network. Data never leaves your environment. That disqualifies cloud-only governance tools in banking and defense, and it is why Daylite fits where they cannot.

Question 6

How long does Daylite take to deploy?

Accepted Answer

Minutes. Docker ships today. Run the container, point your agents at the MCP proxy, and the policy gate and the hash-chain audit are live. A Helm chart for Kubernetes is on the roadmap.

Feature	Daylite (VPC)	Anthropic (Cloud)
Deployment model	Self-hosted single Rust binary inside your VPC	Managed cloud service on Anthropic infrastructure
Data residency	Data never leaves your network. Any region, air-gapped OK	Processed on Anthropic infra, region availability varies
Cryptography	FIPS 140-3 validated library (aws-lc-rs / AWS-LC); not product-certified	Anthropic-managed, no public FIPS certification
Runtime enforcement	Gates each tool-call before it runs: secret scan, PII scan, classification ceiling, per-tool allow/deny/approve	Not disclosed
Audit evidence	Append-only SHA-256 hash-chain + HMAC signatures + legal hold, offline-verifiable (NIST AU-9/AU-10)	Platform-level logging, Anthropic-managed, mutable
PII / PHI handling	Scanner on the request path: regex + Luhn + aho-corasick, auto-redact before the LLM	No built-in PII redaction disclosed
Access control	SCIM 2.0 + OIDC SSO + mTLS + API keys + per-tenant classification ceilings + cell-based multi-tenancy	SCIM RBAC (Apr 9 launch), Workspace roles, SSO via Okta/Entra
Observability	SIEM connector (Splunk HEC, OCSF format), HMAC-signed webhooks with MITRE ATT&CK event types	OpenTelemetry export to Splunk/Datadog (Apr 9 Cowork Enterprise launch)
Standards alignment	NIST SP 800-53 control mapping (AU-9/AU-10, AC-3/AC-4), HIPAA, DORA Article 12, EU AI Act Annex VIII, OWASP Agentic Top 10	SOC 2 Type II (Anthropic corporate), shared responsibility model
LLM provider	Any LLM via OpenAI-compatible API + local inference (candle GGUF). Privacy-aware routing: PII detected = local only	Claude models only. Bedrock option adds AWS FedRAMP High but still cloud-hosted
Secret management	Phantom Tokens: agents see UUIDs, real keys injected at proxy edge. Prevents credential exfiltration via prompt injection	Anthropic-managed credentials
Pricing	One-time deploy, runs on your compute. No per-session fees. Scales with your infra, not your usage	$0.08/session-hour (Managed Agents) + $40-100/seat/mo (Cowork Enterprise)
Network controls	Runs inside VPC, air-gapped support, SSRF protection on all outbound connectors	Internet-accessible SaaS. Bedrock PrivateLink available but Claude-only
Open standards	OpenAI /v1/chat, SCIM 2.0 (RFC 7644), OCSF for SIEM, OIDC, mTLS, MCP JSON-RPC 2.0	Proprietary Messages API. No MCP enforcement layer

Daylite vs cloud AI agent platforms

When to choose Anthropic

When to choose Daylite