Workspace-first AI work

AI workspacesthat teams can trust

Autodeck gives teams integrated Decks for analysis, documentation, grading, and specialist workflows, with the right interface, AI support, and actions already connected.

Workspace

Starts personal, grows into a team without changing the model.

Deck

One integrated surface for a specific job.

Agent role

Assistant, specialist, or worker as the work requires.

Platform principles

Four things the platform already takes seriously

Workspace-first by default

Every user starts in a dedicated workspace. Teams can grow from there without losing history, settings, or agent context.

Governed agents

Assistants, specialists, and workers operate with explicit identity, tenant scope, and auditable tool access.

Apps on one platform

Autodeck keeps multiple app packages on the same foundation, so apps can share the runtime without collapsing into a monolith.

Self-hosted where it matters

The platform is designed for controlled hosting, reproducible environments, and a clear path from local stack to preview and production.

Current Decks

One platform, several distinct integrated surfaces

Read product notes
How a Deck fits together
A Deck is one integrated working surface, not just another page with a chatbot.

Cloud Analyzer Deck

Workspace: ACME Europe

governed
Model choices

Provider selection stays governed instead of being left to guesswork.

Agent role

Specialists and workers run with explicit scope and auditable actions.

Interface

The workflow stays shaped around the job instead of a generic chat window.

Workspace context

History, settings, and access stay consistent as teams expand.

For teams that need structure

Autodeck is not a prompt toy. It is built for teams that need roles, history, approvals, and repeatable behaviour around AI-assisted work.

For product builders

New Decks can live inside the same shell, reuse governed runtime patterns, and keep domain logic in clear boundaries.

For regulated environments

Tenant isolation, governed mutations, and self-hosted infrastructure choices are part of the product model, not an afterthought.

Cloud Analyzer Deck

Inspect Entra ID groups, conditional access, and role intelligence in a tenant-aware workspace surface.

ArchDoc Deck

Draft and maintain architecture documents with shared context, status workflows, and AI-assisted editing.

Toetsie Deck

Run assessment and grading workflows with an app-specific assistant instead of bolting AI onto forms afterwards.

Dungeon Master Deck

A specialist-style surface that exercises the same runtime model in a very different interface, useful as both product and proving ground.

Why Decks matter

Built for real work, not just prompts

Each Deck is shaped around a specific job, with the right interface, AI support, and actions already connected. That makes it faster to use and easier to trust.

01

Start in your own workspace

A new user lands in a personal workspace instead of a shared sandbox.

02

Open the app that fits the job

Each app package keeps its own UI, tools, and specialist behaviour while still using the same platform foundations.

03

Work with the right kind of agent

Assistants advise, specialists drive focused flows, and workers handle background execution.

04

Keep the operating model intact

Tenant boundaries, approvals, and audit trails remain in place as teams, data, and app surfaces expand.

Real product model

See the operating model before you commit

The docs explain workspaces, Decks, and agent roles in plain language. The blog adds product notes on why the product is shaped this way.