Product

Dekaplane is the control plane for self-hosted AIOps.

It gives tenant admins one place to manage model access, AI Gateway routing, governed ChatOps, agent workflows, tool calling, MCP integrations, observability, and operational policy.

Current release Free single-tenant self-hosted

One tenant, local login, packaged install, and current release documentation.

Download free

Built for

Teams that want operational control over AI usage

Self-hosted operators

Run Dekaplane and its supporting services in your own environment while keeping a focused administrative experience.

Platform admins

Centralize model access, provider onboarding, tool enablement, ChatOps configuration, and tenant controls.

AI workflow teams

Use governed ChatOps and agent workflows with approved models, tools, and policy boundaries.

Platform surface

One product surface for gateway, ChatOps, agents, and tools.

Dekaplane is not just a routing screen. It is the administrative control plane where teams configure the model layer, governed workflow behavior, tool access, and tenant-aware operational settings.

Centralized Control Plane Manage provider details, model aliases, routing status, tools, ChatOps access, workflow policy, and traces from one admin surface.
AI Gateway Route local and cloud-backed models through stable aliases with verification and sync visibility.
Governed ChatOps Operate approved ChatOps interactions with tenant-aware access controls, traces, and hosted workflow execution.
Agents + tool calling Support controlled agent workflows that can use approved native tools and MCP-backed integrations.
Observability + debugging Use workflow runs, traces, audit records, and optional telemetry integrations to understand runtime behavior.

Provider onboarding

Bring local and cloud-backed models into one operational model

Local OpenAI-compatible endpoints

Connect vLLM-style or other OpenAI-compatible inference servers and expose them through tenant-facing model aliases.

Azure OpenAI

Add deployment details, API version, and credentials through the same control-plane onboarding flow.

AWS Bedrock

Configure Bedrock model access and region-specific provider details for gateway routing.

Governance

Designed for controlled operations, not ad hoc AI usage

Tenant-aware administration

The current package focuses on one self-hosted tenant with local login, model/provider onboarding, gateway routing controls, tool enablement, and operational visibility.

Controlled runtime workflows

ChatOps and agent flows are designed around approved users, enabled tools, model policy, traceable workflow runs, checkpoints, audit records, and administrative boundaries.

Technical foundations

Built on real workflow, tracing, and gateway foundations

Read the technical notes

Workflow orchestration

Hosted ChatOps and agent flows use checkpointed workflow foundations with durable run and trace records.

Tool governance

Native tools and MCP-backed tools are onboarded, verified, enabled, and exposed through controlled workflow inventory.

Operational visibility

Workflow runs, trace events, logs, metrics, and optional telemetry give operators a practical debugging path.

Today

What single-tenant self-hosted means

One tenant included

The current public package is intentionally simple: one self-hosted tenant in one installed environment.

Local tenant-admin

The installer creates the local admin identity for the deployed product. Portal credentials are not reused inside the install.

Free current package

Sign up for portal access, download the current release, and use the included license and setup guidance.

Get started

Download the current free self-hosted package.

Create a portal account to access the release bundle, license file, release notes, and install guide.