Hubfly spaceDocs
Console

hubfly storage

hubfly storage manages Object Storage: Storage projects, buckets and objects, S3 access keys for your apps, managed assets and video encoding. Uploads and downloads go through short-lived URLs that Hubfly space issues for your account, so you do not need an S3 key to move files from your terminal.

Basics

  • <project> and --project <p> take a Storage project id or name. You can leave it out when you have exactly one Storage project, or set HUBFLY_STORAGE_PROJECT for the shell or CI job. In a terminal the CLI lets you pick when several match.
  • Object paths are written <bucket>/<key>, for example project-assets/reports/hello.txt. Keys are case-sensitive and slashes are part of the key.
  • Storage projects are regional: you choose the region when you create one.
  • Commands that print data accept --json. An S3 secret is printed once, when its key is created.
  • Run hubfly storage help for the full usage text.

terminal

$ hubfly storage create media
$ hubfly storage buckets create project-assets
$ hubfly storage put ./hello.txt project-assets/reports/
$ hubfly storage keys create web-app --bucket project-assets --to web

Storage Projects

usage
hubfly storage list [--org <id|slug>] [--json]
hubfly storage create <name> [--region <id|name>] [--org <id|slug>] [--yes] [--json]
hubfly storage info [<project>] [--json]
hubfly storage usage [<project>] [--json]
hubfly storage delete <project> [--yes]
CommandWhat it does
storage listYour Storage projects with status, region, bucket count, used space and quota, and id. --org narrows the list to one organization.
storage createCreates a Storage project in a region that offers Storage. Pass --region, or pick one in a terminal. --org creates it in an organization instead of your personal account. It shows a summary, asks you to confirm, waits until the project is ready, then prints the S3 endpoint, the quota and the next commands to run. Names are 3–64 characters.
storage infoS3 endpoint, state, quota and used space, bucket count and active video jobs.
storage usageStored bytes against the quota, plus recent daily egress and video encoding minutes.
storage deletePermanently deletes the project with all its buckets, objects, assets and media after you type its name (or pass --yes). Its S3 access keys stop working immediately.

terminal

$ hubfly storage create media --region eu-1 --yes
$ hubfly storage info media
$ export HUBFLY_STORAGE_PROJECT=media # later commands can skip --project

Buckets

usage
hubfly storage buckets [list] [--project <p>] [--json]
hubfly storage buckets create <bucket> [--versioning enabled|suspended|disabled] [--project <p>] [--json]
hubfly storage buckets update <bucket> [--versioning <mode>] [--image-retention <duration>] [--project <p>]
hubfly storage buckets delete <bucket> [--yes] [--project <p>]
CommandWhat it does
buckets listBuckets with their versioning mode and image variant retention.
buckets createCreates a bucket. Names are 3–63 lowercase letters, digits, dots and hyphens, starting and ending with a letter or digit. New buckets start with versioning disabled; --versioning enabled turns it on straight away.
buckets updateChanges versioning (enabled or suspended; once enabled it can be suspended but not disabled) or how long generated image variants are kept (--image-retention, 0 to keep them or 1h to 365d). New buckets start with versioning disabled, and Hubfly space refuses --image-retention on such a bucket on its own: pass --versioning suspended in the same command.
buckets deleteDeletes an empty bucket after you confirm. A bucket that still holds objects is refused; empty it first with storage rm <bucket>/ --recursive.

Read versioning and deletion before turning on versioning: older versions keep using quota.

Objects

usage
hubfly storage ls [<bucket>[/<prefix>]] [--recursive] [--search <text>] [--limit <n>] [--project <p>] [--json]
hubfly storage put <file|dir> <bucket>[/<key>] [--recursive] [--content-type <type>] [--project <p>]
hubfly storage get <bucket>/<key> [<file>|-] [--force] [--project <p>]
hubfly storage rm <bucket>/<key>... [--recursive] [--yes] [--project <p>]
hubfly storage url <bucket>/<key> [--project <p>] [--json]
CommandWhat it does
storage lsWithout an argument, lists buckets. With a bucket or prefix, lists one level like a directory: sub-prefixes first (PRE images/), then objects with size, modification time and key. --recursive lists every key under the prefix, --search filters keys, --limit caps the output. It follows pagination for you.
storage putUploads a file. When the key is missing or ends with /, the file name is appended, so put logo.png assets/img/ stores img/logo.png in the assets bucket. The content type comes from --content-type, else the file extension, else the file contents. --recursive uploads a directory tree under the key prefix, using relative paths with forward slashes and skipping symbolic links: put ./dist assets/site/ --recursive stores dist/a/b.css as site/a/b.css. Files are streamed with progress on standard error; one file can be up to 20 GB.
storage getDownloads an object to a file, by default the last part of the key in the current directory. - writes it to standard output. It does not overwrite an existing file unless you add --force (-f), which the usage line above does not show yet.
storage rmDeletes one or more objects after you confirm. --recursive deletes every object under a prefix: it lists them first and shows how many will be deleted.
storage urlPrints a temporary download URL for an object and when it expires (15 minutes). Anyone with the URL can download the object until then.

terminal

$ hubfly storage ls project-assets/reports/
$ hubfly storage put ./dist project-assets/site/ --recursive
$ hubfly storage put ./export.csv project-assets/exports/2026-10.csv --content-type text/csv
$ hubfly storage get project-assets/reports/hello.txt - | head
$ hubfly storage rm project-assets/tmp/ --recursive

Larger files and multipart uploads

storage put sends each file in one request, up to 20 GB. For larger files, or to resume uploads in parts, use an S3 client with a key from hubfly storage keys create; see the S3 API reference.

S3 Access Keys

usage
hubfly storage keys [list] [--project <p>] [--json]
hubfly storage keys create <name> [--read-only] [--bucket <bucket>] [--format text|dotenv|json]
                                  [--to <cell-project>] [--project <p>]
hubfly storage keys revoke <accessKey> [--yes] [--project <p>]
CommandWhat it does
keys listAccess keys with their name, permission and bucket scope. Secrets are never listed.
keys createCreates a key with read-write access to every bucket, or read-only with --read-only, or limited to one bucket with --bucket. It prints the access key id and the secret once. --format dotenv or json prints them as the variables below.
keys create --toWrites the variables into the environment of a Cell project instead, with the secret stored as a secret variable, and prints only the variable names. The project's containers are redeployed to load them. If writing the variables fails, the new key is revoked.
keys revokeRevokes a key after you confirm. Requests signed with it fail from then on.

The variables written by --format dotenv and --to:

VariableValue
HUBFLY_STORAGE_ENDPOINT, AWS_ENDPOINT_URL_S3The project's S3 endpoint
HUBFLY_STORAGE_REGION, AWS_REGIONThe signing region, us-east-1 unless the Dashboard reports another. Hubfly space Storage accepts any SigV4 signing region.
HUBFLY_STORAGE_ACCESS_KEY_ID, AWS_ACCESS_KEY_IDThe access key id
HUBFLY_STORAGE_SECRET_ACCESS_KEY, AWS_SECRET_ACCESS_KEYThe secret (secret variable)
HUBFLY_STORAGE_BUCKETThe bucket, only for bucket-scoped keys

Recent AWS SDKs read the AWS_* names without extra configuration. Turn on path-style addressing in your S3 client.

terminal

$ hubfly storage keys create web-app --bucket project-assets --to web
$ hubfly storage keys create backup-reader --read-only --format dotenv > .env.storage
$ hubfly storage keys revoke <accessKey>

Assets And Video

usage
hubfly storage assets [list] [--bucket <bucket>] [--project <p>] [--json]
hubfly storage publish <bucket>/<key> [--protected] [--title <text>] [--project <p>] [--json]
hubfly storage share <bucket>/<key>|<assetId> [--expires <duration>] [--project <p>] [--json]
hubfly storage unpublish <assetId|bucket/key> [--yes] [--project <p>]
hubfly storage encode <bucket>/<key> [--max-height 360|480|720|1080] [--wait] [--project <p>] [--json]
hubfly storage media [list] [--project <p>] [--json]
CommandWhat it does
assets listManaged assets, optionally for one bucket, with their access mode and delivery URL.
storage publishRegisters an uploaded object as a managed asset and prints its id and delivery URL. Assets are public unless you pass --protected, in which case they are served only through share links.
storage shareCreates a share link for an asset that expires after --expires (5 minutes to 24 hours, default 1 hour). For video it returns a player URL and an HLS manifest URL.
storage unpublishDeletes the managed asset after you confirm. Deleting an asset also deletes its source object, renditions and image variants, and breaks links that use it.
storage encodeStarts HLS encoding of a video object up to --max-height (default 720). Encoding runs in the background; --wait follows the job until it is ready or fails.
storage mediaVideo encoding jobs with their state, progress and errors.

terminal

$ hubfly storage put ./logo.png project-assets/images/
$ hubfly storage publish project-assets/images/logo.png --title 'Logo'
$ hubfly storage put ./intro.mp4 project-assets/videos/
$ hubfly storage encode project-assets/videos/intro.mp4 --max-height 720 --wait
$ hubfly storage share project-assets/videos/intro.mp4 --expires 2h

Scripting And Secrets

  • JSON output. With --json, standard output carries only JSON; prompts and progress go to standard error.
  • Confirmations. Creating a project and deleting anything ask first. Without a terminal they stop with confirmation required; re-run with --yes; pass --yes once you have reviewed the change. Deleting a project asks you to type its name.
  • Secrets. An S3 secret is printed once, when the key is created, or written straight into a Cell project with --to. Store it at once; Hubfly space cannot show it again. Temporary upload, download and share URLs are also credentials until they expire. --debug never prints secrets.
  • Default project. Set HUBFLY_STORAGE_PROJECT in CI so commands do not depend on how many Storage projects the account has.
  • Exit codes. Non-zero on failure, with distinct codes for usage errors, authentication, conflicts and timeouts. See exit codes.

Upload a build in CI

Set HUBFLY_TOKEN and HUBFLY_STORAGE_PROJECT, then run hubfly storage put ./dist project-assets/site/ --recursive. Your app itself should use an S3 key from keys create, not your Hubfly space token.
Something unclear or out of date? Emailsupport@hubfly.spaceBack to top