You run your business. We run the platform.
DeckCrew is fully managed. No servers to provision, no containers to maintain, no GPU costs to manage. Your agents run in isolated workspaces with automatic scaling and updates.
Fully managed, fully isolated
Every company gets its own workspace. Your agents, knowledge base, and configurations are completely separate from every other team on the platform.
- No servers to manage We handle hosting, scaling, and availability.
- Workspace isolation Your data never touches another company's workspace.
- Automatic updates Always on the latest version without downtime.
- Regional deployment Choose where your workspace runs.
Per-workspace runtime isolation, engine-managed agent wakeup, centralized scheduling and approval infrastructure. Worker agents scale to zero and wake on dispatch.
The engine runs everything behind the scenes
Message routing, agent coordination, scheduled tasks, approval pipelines, and channel normalization — all managed by the DeckCrew engine so you never think about infrastructure.
- Message routing The right message reaches the right agent automatically.
- Agent scaling Agents wake up when needed and sleep when idle.
- Approval pipeline Draft-and-approve flows run without manual wiring.
- Channel normalization Slack, Gmail, API, and Web Chat all work the same way internally.
Engine-managed dispatch, channel abstraction layer, workflow step contracts, and centralized approval read/write APIs.
Your keys, managed securely
Connect your own OpenAI and Claude keys. Billing stays with your provider, while DeckCrew manages workspace-scoped credential access securely for the hosted platform.
- BYOK by design Provider billing stays in your own accounts.
- Per-agent provider scoping Different agents can use different keys and models.
- Connection management Add, remove, and rotate keys from the dashboard.
Provider-connection admin API with OAuth flows, per-agent provider/tool policy layering, and encrypted credential storage.
Platform FAQ