crewship
Start herenewYour infrastructureAI & memoryTeam workspaceAccess & oversightChangelogDocs
GitHub Discord
Install
$ docker run -p 8080:8080 crewship/crewship:latest
YOUR AI. YOUR KNOWLEDGE.

Choose the engine. Keep the context.

Use the AI that fits the work. Give your agents persistent context without tying your whole workspace to one provider.

01

Choose a supported adapter and model.

Connect your own provider credentials through a supported CLI adapter. Model selection and authentication depend on that adapter. Your workspace can use different configured engines for different work.

See CLI adapters
02

Run local inference with Ollama.

Configure a supported local model setup when inference needs to stay on your infrastructure. Hardware requirements, tool support, and output quality depend on the model you choose.

Explore local models
03

Let useful knowledge carry forward.

Enable persistent agent memory and deliberately scope shared crew knowledge. Keep procedures and decisions available across sessions, so colleagues can work with the same specialist and its context.

Understand memory

A specialist your team can return to.

  1. A colleague explains an approved reconciliation procedure.
  2. With memory enabled, relevant context remains available for later work.
  3. Another colleague with appropriate access can continue with the same specialist.

Stored memory stays on your host. A cloud model receives the context included in its requests; use a local model and controlled tools when data must remain on-premises.

INSIDE CREWSHIP

One platform.
See the real workspace.

Actual screenshots from our development instance, with demo data.

Meet the agent behind the work.

Riley’s profile brings together run outcomes, conversations, assigned work, and connected resources.

Captured from dev1 · Demo workspace · September 2026
Riley’s profile brings together run outcomes, conversations, assigned work, and connected resources.
START WITH ONE CREW

Make room for your first task.

Get started