DreamLake

Quick start

From login to a verified round-trip in a few commands.

Step 1 — Log in

The CLI ships with built-in environments and logs into prod by default:

terminalbash
dreamlake login

A browser opens for device authorization. Approve it, and your token is saved. Confirm:

terminalbash
dreamlake profile

It prints your user and namespace — note the namespace, you'll use it in targets below.

Step 2 — Create a project

terminalbash
dreamlake create project my-robots --description "robot captures"

Projects are private by default. Pass --public to share. The output shows the project slug — use it in targets.

Step 3 — Upload a file

terminalbash
dreamlake upload ./clip.mp4 --episode my-robots:run-001 --to /camera/front
  • my-robots:run-001 is a target: project[:episode].
  • The episode run-001 is created automatically if it doesn't exist.
  • --to is the path inside the episode.

Step 4 — List and download

terminalbash
dreamlake list --episode my-robots:run-001
dreamlake download --episode my-robots:run-001 --from /camera/front/clip.mp4 -o ./out.mp4
That's the full loop

Create → upload → list → download. From here, explore uploading (folders, resume), projects & data (bindrs, datasets), or organizations.

Targets at a glance

Most data commands take a target string:

FormUsed byExample
project[@namespace]--projectmy-robots@alice
project[@namespace][:episode]--episodemy-robots@alice:run-001

@namespace is optional — it defaults to your own (from dreamlake profile).