Your agents built to endure.

Securely run agents you don’t trust.

Agents are easy to demo and hard to trust. Andyur runs them anyway. It works out exactly what one run may do, then enforces that outside the agent. The agent holds no production credential and is offered no tool beyond its grant.

Bring the container you already have. No SDK.

agent manifest requestresolution granted
runtime:
  image:
    ref: ghcr.io/acme/agent
    digest: sha256:8b2c…5fbd
  lifecycle:
    max_seconds: 7200
model:
  requested: claude-sonnet
  access: proxy
capabilities:
  tools:
    - server: tickets
      tools: [read_ticket,
              close_ticket]
image     sha256:8b2c…signed
command   as reviewed

lifetime  7200 3600clamped

model     claude-sonnet
api key   at the brokerbrokered

tools
  read_ticketgranted
  close_ticketdropped

secrets   none in the run

How a run is governed

Decided in one place. Enforced outside the agent.

Every run passes through the same three steps, and enforcement never depends on the agent behaving.

01

Declare

A manifest says what the agent wants: image, command, model, tools, lifetime. It grants nothing.

02

Resolve

Policy intersects the request with what the platform allows. The result can only get smaller. It is signed, and pinned to the image digest and command that were reviewed.

03

Govern

Each run gets its own gateway. It strips whatever the agent set, attaches the approved credential, and routes only to the approved target.

Identity

Named

Every agent and every run gets a cryptographic identity. Short-lived, attested, verifiable at every hop. No shared API keys, and no vendor credential inside the agent.

Governance

Bounded

A manifest is a request, never a grant. What one run may do is computed once, server-side, from the acting user, the agent's ceiling, and the target. It expires with the run.

Observability

Witnessed

One run, one trace. Every decision, tool call, and handoff recorded and attributable to the agent that made it, across every process it touched.

Bring your own agent

Your container. Your framework. Our boundary.

Andyur runs the image you already build, on any agent framework or none. There is no SDK to adopt and nothing to import. The boundary sits outside the container, so the agent does not need to know Andyur exists.

Next: unmodified stock open-source tools, run under the same boundary with no Andyur awareness at all.

  • Self-hostedDocker on one host, or Kubernetes for the cluster. Your infrastructure, your data.
  • Nothing hiddenEvery hardening control is named, and the platform refuses to start in production without it.
  • Nothing forcedIdentity plane, policy engine, credential vault and tracing are each opt-in. The first run needs none of them.

Built on open standards

Composed, not reinvented.

Every seam is documented, so you can swap in your own identity provider, policy engine, or vault, and it never locks you in.

SPIFFE / SPIRE OAuth Token Exchange OpenID Connect Open Policy Agent OpenTelemetry Envoy OpenBao MCP