platform components, fully documented
native instrumentation throughout
typical integration timeline
SaaS deployments live in your tenancy
The documentation is structured so that each audience can land in the right index without scanning the rest. The links below are placeholders during the design partner phase. Full content ships publicly with the 2027 GA release.
Reference architectures for SaaS-in-VPC, single-tenant on-premises, and air-gapped deployments. Identity integration patterns. Network topology. Failure modes and reversibility model.
Sentinel SDK installation and quick start, Gateway integration, MCP registry management, policy authoring, and OTel instrumentation. Worked examples in Python, Node, and Go.
Threat model, detection coverage matrix, NPI/PCI/PII redaction classes, Kill Switch invocation paths, and incident response runbooks. SOC 2, pen test, and subprocessor artifacts.
Regulatory mapping for GLBA / Reg P, PCI-DSS 4.0, EU AI Act, DORA, GDPR, CCPA, and HIPAA. Pre-completed VRA, SIG, and CAIQ packages. Audit trails and evidence chain documentation.
The full quick start is in the design partner doc set. The four-step shape below is what every onboarding looks like, same sequence whether you are pilot or production.
Email contact@bladerun.ai with a one-paragraph description of your AI agent fleet, your model providers, and your target deployment topology. A founder responds within one business day.
Working session with the founders and your architecture team. Output is a written scope memo: deployment shape, integration plan, identity model, and the first workflow to govern.
BladeRun engineering provisions the pilot environment in your VPC or on your hardware. Gateway, Sentinel SDK packages, MCP Server Registry, and Time Machine storage configured for your account.
Install the Sentinel SDK in your agent runtime, Python, Node, or Go. Point one outbound LLM call through the Gateway. Inspect the first signed agent traffic in Time Machine within minutes.
Both reference sets are gated to design partners under NDA today. Each card describes what is in the reference and how to get access.
Full REST and gRPC reference for the Gateway, MCP Server Registry, policy management, Time Machine queries, and the Kill Switch invocation path. OpenAPI 3.1 spec and machine-readable schemas.
Per-language reference for the Sentinel SDK, agent identity attestation, tool-call inspection hooks, OTel instrumentation, and policy enforcement primitives. Full type signatures and worked examples.
Documentation is currently under design partner NDA. The full doc set, SDK reference, API reference, runbooks, and incident response procedures are available to design partners and active enterprise prospects in legal review. Send a one-paragraph note describing your role and your evaluation context to contact@bladerun.ai and we will provision access on signed mutual NDA, typically within one business day.
Public documentation ships with the 2027 enterprise GA release. The cadence is intentional: during the 2026 design partner phase, the doc set evolves alongside the platform with input from the founding cohort. Freezing public docs before the design partner cycle ends would create drift between what the cohort sees and what new readers see. We would rather keep the public surface honest and ship it once.
Public today: the platform architecture, the Federation Network white paper, the threat taxonomy, deployment topology overviews, and the buying motion (see /pricing). NDA-gated today: the SDK reference, full API reference, runbooks, SOC 2 Type I report, pen test results, full subprocessor list, and the deployment-specific reference architectures. The split mirrors how Tier 1 banks already triage vendor materials.
Yes, under NDA. Worked examples for the Sentinel SDK in Python, Node, and Go are shared during the technical session, including the agent identity attestation pattern, the tool-call inspection hook, and a complete OTel instrumentation example wired into a typical bank agent. Sample policy bundles (YAML and OPA) are shared on the same path.
A shared Slack or Teams channel is provisioned with the founders and the BladeRun engineering team. Pilot SLAs are documented in writing. Typical response is under one business day for engineering questions and under four hours for any pilot-blocking issue. After GA the named TAM model takes over; founder access remains available on every tier.
Send a one-paragraph note describing your role and evaluation context. Mutual NDA in one pass, doc access provisioned within a business day.