Open source SDK + managed platform

Control what your AI agents can do.

AI agents have unrestricted access to your APIs, databases, and infrastructure. Sekuire enforces guardrails at runtime - before damage happens. Works with any framework, model, or cloud.

Free forever. No credit card required. Set up in under 5 minutes.

sekuire verify
$ sekuire verify --agent billing-agent
Identity: 8573fa8a...732fced
Signature: Ed25519 valid
Policy: 3 rules loaded
BLOCKED tool.http.request to *.corp.internal
Identity
BLAKE3 + Ed25519
Enforcement
Local, offline-first
Last 24 hoursAll agents compliant
847
allowed
23
blocked
5
flagged

Works with your stack

🔗LangChain
🤖CrewAI
MCP
Vercel AI SDK
AWS
Azure

The problem

AI agents run with full permissions by default

Most AI agents today have unrestricted access to your tools and infrastructure. This creates risk that scales with every agent you deploy.

Call any API without restriction
Send emails on your behalf
Read and modify databases
Execute arbitrary code
Rack up costs with no limit
Without Sekuire
High risk
  • Shadow agents running without oversight
  • No way to enforce behavior constraints
  • Audit logs scattered across systems
  • No kill switch for compromised agents
With Sekuire
Enforced
  • Full fleet visibility in real time
  • Runtime policy enforcement on every action
  • Immutable, centralized audit trails
  • Instant credential revocation

Three lines to govern your agents

Install the SDK, define your policy, and enforce it at runtime. No account required.

1

Install

Add the SDK to your agent project. Works offline with no account required.

npm install @sekuire/sdk
2

Configure

Define behavior constraints in sekuire.yml with policy-as-code.

sekuire init --name my-agent --language typescript
3

Enforce

Runtime checks block unsafe actions before execution.

const allowed = await sdk.checkPolicy(action, context)

See what your agents are doing

Agent inventory, policy decisions, and containment events in one place.

Dashboard Preview
Live governance operations workspace
Request Demo
Active Agents
142
12 deployed today
Policy Checks
99.2%
Pass rate (24h)
Pending Reviews
7
3 high priority
Kill Switch Events
0
Last 30 days

Fleet Status

Agent health and compliance overview

View All
AgentStatusLast Heartbeat
Sales Assistant Pro
active
2 mins ago
Data Pipeline Bot
active
15 mins ago
Analytics Engine
warning
5 mins ago

Activity

Recent events

New agent 'Invoice Processor' deployed

10 minutes ago

Verification completed for 'Sales Bot v2.1'

25 minutes ago

Unusual activity in 'Analytics Engine'

1 hour ago

Open SDK

Free forever

Local policy enforcement, identity hashing, and audit sinks. No account required.

  • Policy-as-code with "sekuire.yml"
  • Runtime enforcement (offline-first)
  • Identity hashing and signatures
  • Local audit sinks (console/file)
  • Works with any framework or model
typescript
import { SekuireSDK } from "@sekuire/sdk"

const sdk = SekuireSDK.fromEnv()
await sdk.start()

const allowed = await sdk.checkPolicy(
  "tool.http.request",
  { url: "https://example.com" }
)
if (!allowed) throw new Error("Blocked")

Sekuire Cloud

Managed

Fleet dashboard, centralized policy distribution, registry publishing, and compliance reporting.

  • Fleet dashboard and inventory
  • Centralized policy distribution
  • Registry publishing and verification
  • Immutable audit retention
  • SSO and RBAC (Enterprise)

Your agents have identity. Now give them rules.

Identity tells you who an agent is. Sekuire tells it what it can do. Runtime policy enforcement, cryptographic audit trails, and instant containment - built for security teams who need to prove compliance, not just hope for it.

Stop shadow agents

See every agent running, where it is deployed, and who owns it.

Prevent unsafe actions

Runtime enforcement blocks risky tool calls before they execute.

Prove compliance

Audit trails built for SOC2 and internal reviews.

AGENT INVENTORY
6 agents discovered
Sales_Assistant_v3
AWS us-east-1
active
2s ago
Data_Pipeline_Bot
GCP europe-west1
active
5s ago
Customer_Support_AI
Azure westus2
active
12s ago
Unknown_Crawler_X
Unknown
critical
?
Analytics_Engine_v2
AWS us-west-2
warning
45s ago
Invoice_Processor
On-Prem DC1
active
3s ago

Built for your team

Different roles, same need: visibility and control over autonomous agents.

CISOs

  • Full agent inventory
  • Kill switch capability
  • Compliance-ready audit logs

Platform Engineers

  • Beacon auto-discovery
  • Policy-as-code workflow
  • CI/CD integration

Compliance Officers

  • SOC2/HIPAA reports
  • Immutable audit trail
  • Delegation tracing

Production controls for your agents

Permissions, enforcement, and audit. Not just visibility.

Runtime Policy Enforcement

Define what agents can and cannot do. Sekuire enforces policies during execution and blocks unsafe actions before they occur.

Define rules in a simple YAML file. The SDK enforces them locally at runtime - no network calls, no latency, no single point of failure.

sekuire.ymlactive
rules:
  - action: tool.http.request
    effect: deny
    when:
      url_match: "*.corp.example.com*"

  - action: tool.database.write
    effect: deny
    when:
      table: "users"
      operation: "DELETE"

Live Agent Inventory

See every AI agent running across your organization in real time, including deployment location, health, and identity.

Task Delegation Tracing

Trace how agents delegate work to each other, including execution paths, failures, and timing.

Audit and Compliance

Automatically generate immutable logs designed for security reviews and compliance requirements.

Kill Switch

Instantly revoke an agent's credentials to stop it from acting.

Security-first architecture

Built for security teams

Ed25519 signatures

Every agent action is cryptographically signed and verifiable.

Immutable audit logs

Tamper-proof records of every policy decision and tool invocation.

Offline-first enforcement

Policies enforced locally. No network dependency, no single point of failure.

OAGS open standard

Built on the Open Agent Governance Specification. No vendor lock-in.

"We went from 'we think our agents are fine' to actually proving it. The policy enforcement caught three misconfigured agents in the first week."
Early design partner - Platform Engineering Lead

Start free, scale when you need to

Start free with the open SDK. Upgrade when you need fleet visibility and centralized control.

Free
$0

Open SDK. Offline-first enforcement.

Starter
$1,000/mo

25 managed agents. Registry publishing. 30-day retention.

Most popular
Professional
$5,000/mo

100 managed agents. Full registry. SIEM integrations. 90-day retention.

Enterprise
Custom

Unlimited agents. SSO/RBAC. Custom retention. Private deployment options.

Questions about getting started?

Deploy guardrails now

Three lines of code. No account required. Start enforcing guardrails in under 5 minutes.