ReticleMake the diagram you'd make anyway
Arrange services, boundaries, dependencies, and notes freely in a visual editor built for infrastructure.
YOUR ARCHITECTURE. KEPT LIVE.
The infrastructure diagram you can operate.
Reticle starts with the infrastructure diagram your team would make anyway. Map your team's mental model, then connect it to live systems. See current health, investigate problems, and act in context.
The same read-only context is available to teammates, internal tools, and agents through the browser, JSON, and MCP.
One YAML · Visual editing · Customer-hosted · No Reticle cloud required
A REAL DIAGRAM, KEPT LIVE
See a hand-authored architecture diagram connected to real systems.
This is the map we use to understand the infrastructure serving reticle.live. We drew it to express that mental model, then connected its services to continuous health evidence from the Team Daemon it describes.
Pan, inspect, and export the map. The public view is read-only. Open full screen ↗
DIAGRAMMING FIRST
Draw the architecture as your team understands it. Then attach reality.
The best infrastructure diagrams are made intentionally: they leave out noise, show the boundaries that matter, and explain relationships. Reticle gives that work a purpose-built canvas, then makes the result live.
Express intent, not inventory
Choose the level of detail your team uses to reason about the system instead of accepting a generated view.
Connect it to reality
Attach current health evidence and guarded actions directly to that model, then share it with people and tools.
DESKTOP AND TEAM
Start local. Share when it matters.
Start with the free Desktop app.
Free, MIT-licensed, and local-only
Map a homelab or production-shaped environment locally. Keep each diagram in Git, attach live checks, and export or query it without creating an account or sending topology to a Reticle service. Desktop supports multiple local workspaces.
Keep the diagram live for the whole team.
Customer-hosted and always on
Run one daemon inside your environment. It continuously checks one shared topology and serves it to authorized browsers, internal integrations, and read-only agents. The diagram stays live when one operator closes a laptop.
FOR DAILY WORK AND INCIDENTS
Understand the system before changing it.
See the system around the failure.
See the affected service, its dependencies, current evidence, notes, and freshness without reconstructing the architecture aloud.
Keep architecture outside one person's head.
Give new operators and escalation paths the same system model instead of relying on the person who originally built it.
Put evidence beside the response.
Keep current evidence beside bounded, preconfigured actions with approvals, timeouts, and optional preconditions.
INSPECTION BEFORE EXECUTION
Context for agents. Control for operators.
The visual UI, JSON API, and MCP inspect the same current diagram. MCP remains read-only, Team exposes no browser shell, and guarded actions stay explicit human workflows.
Read-only machine access
MCP has no shell, mutation, command, or named-action tools. Optional chat is another read-only consumer, not the product.
Explicit operations
Named actions are persisted server-owned definitions with approvals, preconditions, and timeouts. Callers never submit command text.
Deliberate trust
Custom command checks are arbitrary commands and require privileged configuration. The OS identity and restricted SSH principal remain the execution boundary.
Topology context raised patch correctness from 11.1% to 78.0% in controlled Kubernetes trials ↗
TEAM PRICING
One topology. Unlimited teammates.
Each subscription covers one customer-hosted Team Daemon and one shared topology. Team adds an always-on network vantage point, server-enforced viewer/editor access, and configurable JSONL audit logging.
$199/month
Per running Team Daemon, billed monthly.
$1,999/year
Per running Team Daemon, billed annually.
$3,000one-time
We deploy Team, establish TLS and access controls, and turn the agreed architecture from your diagrams, runbooks, and operator knowledge into the first live topology.
Start with Team
We reply within one business day. Card, ACH, or bank transfer.
Deployment boundaries
One daemon serves one topology. Access uses shared viewer and editor bearer tokens; SSO, individual identity, per-user attribution, and a default SLA are not included. Audit logging is configurable, not automatic.