CLI
Commands
Look up every staqly command and its flags, from running the server to posts, channels, schedules, the queue, alerts and connecting agents.
This page lists every staqly command with its flags and an example. Use it when you know what
you want to do and need the exact syntax. New to the CLI? Start with the
CLI overview, which covers installing and connecting to a server.
Run staqly <command> --help for the full flag list with examples. Wherever a command takes a
collection, you can pass its name or its ID; a channel can be given by name too.
Every command also accepts the global flags --api-url (alias --url), --key, --profile,
--json, --no-color and --help; see Connect to a server.
Most command groups have short aliases:
| Command | Aliases |
|---|---|
posts | post |
channels | channel |
schedules | sched |
collections | collection, cols |
alerts | alert |
jobs | queue |
media | files |
webhooks | webhook, hooks |
workspace | ws |
plans | plan |
keys | key, api-keys |
sessions | session |
profiles | profile |
A typical session
Check the connection
staqly statusShows server health and version, mode, workspace, who you are, the masked key, your plan, usage against limits and how many posts are scheduled.
Connect a channel and write a post
staqly channels connect custom --name "My publisher"
staqly posts create "Product updates" --content @post.md --channel "My publisher" --publish-at "2026-11-03 09:00"Follow it
staqly schedule
staqly posts get post_oalirbz67y4m4cqxq5dwRun the server: serve
staqly serve [--data DIR] [--port 3333] [--dashboard-port 8080] [--expose]
staqly serve token [--rotate] | connect-link | backup <file> | storage migrateRuns the API, MCP endpoint, job queue and dashboard in one process, without Docker. Data lives
in ~/.local/share/staqly unless you pass --data (or set STAQLY_DATA_DIR). --expose opens
the dashboard to your network. Server environment variables such as LOG_LEVEL,
STAQLY_SECRET_KEY and PUBLIC_DASHBOARD_URL still apply; serve defaults PUBLIC_API_URL to
http://localhost:<port> and LOG_FORMAT to text. At every start it prints the access token
and a http://localhost:<dashboard-port>/#token=... link that signs the browser in.
| Flag | Type | Required | Description |
|---|---|---|---|
--data | path | No | Folder for the SQLite database, uploads, secret.key and access_token (default ~/.local/share/staqly, or STAQLY_DATA_DIR) |
--port | integer | No | API and MCP port, always bound to 127.0.0.1 (default 3333) |
--dashboard-port | integer | No | Dashboard port (default 8080) |
--expose | boolean | No | Serve the dashboard on all network interfaces; phones sign in once with a QR code or a link |
With a subcommand it runs an operator command against the same data instead, and exits:
| Subcommand | What it does |
|---|---|
token [--rotate] | Prints the access token (stdout: only the token) and, on stderr, a link that signs a browser in; --rotate makes a new one, the old one stops working and every browser is signed out |
connect-link | Prints a one-time sign-in link (5 minutes, single use) |
backup <file> | Writes a consistent copy of the database while the server runs |
storage migrate [--from DIR] [--dry-run] | Copies media files into the S3 bucket of STORAGE_DRIVER=s3 (Media storage) |
See Run the binary.
Run it in the background: service
staqly service install [--data DIR] [--port 3333] [--dashboard-port 8080] [--expose]
staqly service status | logs [-f] | uninstallInstalls staqly serve as a background service of your user (macOS LaunchAgent, started at
login; Linux systemd user unit, started at boot), restarted if it stops. Environment settings
(STAQLY_*, PROVIDER_*, PUBLIC_*, LOG_*...) at install time are kept. Running install
again replaces it. See Run it in the background.
Check the connection: status
staqly status # health + version, mode, workspace, principal, masked key, plan, usage vs limits, scheduled count
staqly status --jsonWrite and publish: posts
staqly posts list [--collection C] [--status draft|scheduled|publishing|published|failed] [--channel CH] [--tag t] [-q text] [--publish-after 2026-10-20] [--publish-before 2026-10-27] [--sort publish_asc] [--limit 50] [--cursor X] [--all]
staqly posts get post_123 # content, media, targets with their status and link, history summary
staqly posts create <collection> [--content TEXT|-|@file --stdin] [--channel CH ...] [--media med_1 ...] [--file PATH ...] [--tag t ...] [--metadata '{json}' --meta k=v] [--source s] [--external-ref r] [--repeat 1w] [--reply-delay 5m --reply TEXT ...] [--publish-at TIME | --now] [--wait [--timeout 10m]]
staqly posts update post_1 [--content TEXT] [--channel CH ...] [--media ...] [--tag ...] [--metadata '{json}'] [--collection C] [--publish-at TIME | --clear-publish-at] [--repeat 1w|none] [--reply TEXT ... | --no-replies] [--no-channels] [--expected-version N]
staqly posts schedule post_1 <time> # set the time, status scheduled
staqly posts publish post_1 [--wait [--timeout 10m]]
staqly posts cancel post_1 # off the calendar, back to draft
staqly posts duplicate post_1 # a new draft with the same content, media, tags and channels
staqly posts preview post_1 # the text per channel (links removed or defanged) and what each platform objects to
staqly posts stats post_1 --refresh # its numbers per channel (views, likes...) where the platform reports them
staqly posts share post_1 --expires 7 # a preview link anyone can open and comment on (shown once)
staqly posts comments post_1 # what reviewers said on the preview
staqly posts retry post_1 <channel> [--wait] # one failed channel again
staqly posts result post_1 <channel> --status published|failed [--url https://...] [--external-id 1790] [--error "<text>"]
staqly posts history post_1 # events, oldest first
staqly posts delete post_1 [post_2 post_3] # history is kept; API keys need posts:delete
staqly posts purge post_1 [--yes] # post, its events and unused media (admin)
staqly posts tags # the workspace's tags, most used first
staqly work # what your automation is asked for: targets waiting for a result, schedules asking for postsCreating posts
| Flag | What it does |
|---|---|
--content TEXT | The text. --content @path reads a file; --content - or --stdin reads stdin |
--channel CH | A target on this channel (name or ID), comma-separated or repeated. A channel's own text and provider settings are set in the dashboard, over REST or MCP; a channel's default settings come from channels update --setting key=value |
--media med_... | A file already uploaded with media upload, in order, repeatable |
--file PATH | Uploads a file and attaches it, repeatable |
--tag | A label, repeatable (up to 20) |
--metadata '{json}' | Metadata as a JSON object |
--meta k=v | One metadata key, repeatable. Values that parse as JSON keep their type: seed=42 is a number, tags=["a"] an array |
--source | Who wrote it (default cli) |
--external-ref | Your own reference, e.g. a campaign or ticket id |
--repeat | Bring the post back after each publish: 1d to 6d, 1w, 2w, 1mo; none (on posts update) stops it. posts get shows it as Repeats every week |
--reply TEXT | A reply published under the post (a thread part), repeatable, in order. On posts update the replies given replace the thread; --no-replies removes it. posts get lists the thread with each channel's reply state |
--reply-delay 5m | How long the next --reply waits after the part before it (0 to 24h); applies to that one reply |
--publish-at TIME | Schedule it: RFC 3339 (2026-11-03T09:00:00Z), a local day or time (2026-11-03, 2026-11-03 09:00) or relative (+90m, +2h, +3d, +1w). Needs at least one channel |
--now | Publish right away (needs at least one channel) |
--wait, --timeout | With --now or --publish-at: wait until the post is published or failed, like posts publish --wait (default 10m) |
--idempotency-key | Makes retries safe: the same key returns the same post. JSON creates send an automatic key when you don't set one |
Publishing and results
schedule,publish,cancel,retryandresultneed theposts:publishscope;create,updateandduplicateneedposts:write.--waitoncreate,publishandretrypolls the post until it ispublished(exit0) orfailed(exit2), or--timeoutpasses (exit3, default10m). See exit codes.resultreports what your own automation did with a target:publishedwith--urland--external-id, orfailedwith--error.update --channelreplaces the whole target list (--no-channelsremoves every target);--expected-version Nfails with409unless the post is still at that version. A published post cannot be updated:duplicateit.deleteremoves posts but keeps their events in the activity log.purgealso deletes every event about the post, their webhook deliveries and its media when no other post uses them. It needs theadminscope and asks you to type the post ID unless you pass--yes(without--yesit refuses when stdin is not a terminal).
The calendar: schedule
staqly schedule # the next 7 days: scheduled posts and upcoming schedule runs
staqly schedule --days 30 --collection "Product updates"
staqly schedule --all # every post with a time in any status, including past ones
staqly schedule --json # the raw post page (sort publish_asc)Posts with a time, soonest first, with the schedules' next runs between them. Set a time with
--publish-at on posts create or posts update, or with posts schedule.
Connect accounts: channels
staqly channels providers # the catalog: id, name, auth, status, limits
staqly channels providers telegram # one provider with its credential and settings fields
staqly channels list [--group "Client A"] # with the group and last post; --group keeps one group
staqly channels get ch_123 # or by name, with its group, last and next post; credentials are never shown
staqly channels connect <provider> [--name <text>] [--group <name|id>] [--instance <host>] [--field key=value ...] [--fields-json '{json}'|-|@file] [--setting key=value ...]
staqly channels update ch_123 [--name <text>] [--group <name|id>] [--setting key=value ...] [--links keep|remove|defang] [--enable | --disable] [--field key=value ...]
staqly channels test ch_123 # exit 1 when the platform refuses the stored credentials
staqly channels remove ch_123 # at once, no confirmation; targets on it go with it
staqly channels app x --client-id ... --client-secret - # this workspace's own developer app for an OAuth platform; --remove drops it
staqly channels groups list # channel groups with their channel counts; Default always exists
staqly channels groups create "Client A" [--color "#22c55e"]
staqly channels groups rename "Client A" "Client Alpha"
staqly channels groups remove "Client A" # its channels move to Default
staqly channels links create [--note <text>] [--platforms x,bluesky] [--group <id|name>] [--hours 168] [--uses 1]
# a connect link for someone without access; printed once
staqly channels links list # status (open, used, expired, revoked) and uses
staqly channels links revoke chl_123 # channels already connected through it stay| Flag | Type | Required | Description |
|---|---|---|---|
<provider> | string | Yes | connect: a provider id from channels providers with auth token or none (custom). OAuth platforms connect in the dashboard |
--name | string | No | What to call the channel; defaults to the account's name |
--field | key=value | No | One credential field, repeatable (bot_token=..., chat_id=...). On update, fresh credentials to reconnect with |
--fields-json | JSON | No | The credential fields as a JSON object; - reads stdin and @path a file, which keeps secrets out of your shell history |
--setting | key=value | No | A provider-level default for new targets, repeatable |
--enable, --disable | boolean | No | update only: turn the channel on or off |
--group | string | No | A channel group by name or ID. list: only that group. connect: put the channel there. update: move it there |
--instance | string | No | connect: for platforms with many servers (Mastodon), the server's URL or host |
--client-id, --client-secret | string | No | app: the workspace's own OAuth developer app, used instead of the server's PROVIDER_<ID>_CLIENT_ID / _CLIENT_SECRET; --client-secret - reads stdin. See Developer apps |
Managing channels needs channels:manage; listing them channels:read. See
Channels.
Publish on a rhythm: schedules
staqly schedules list [--collection C]
staqly schedules get sch_123
staqly schedules create <collection> --name "Weekday mornings" --times 09:00,17:00 [--weekdays mon,tue,wed,thu,fri] [--tz Europe/Warsaw] --channel CH [--pick next|all|selected --post post_1 ...] [--source none|agent|endpoint --endpoint-url https://... --endpoint-secret s] [--disabled]
staqly schedules create <collection> --name "Every 6 hours" --every 6h --channel CH
staqly schedules create <collection> --name "Launch day" --at "2026-11-06 18:00" --channel CH --pick selected --post post_1 --post post_2
staqly schedules update sch_123 [--name] [--at ... | --every ... | --times ...] [--weekdays] [--tz] [--channel ...] [--pick] [--post ...] [--source] [--endpoint-url] [--endpoint-secret] [--enable | --disable]
staqly schedules run sch_123 # now; the regular next run is unchanged
staqly schedules delete sch_123| Flag | Type | Required | Description |
|---|---|---|---|
--name | string | Yes | Shown on the calendar's hollow chips |
--times HH:MM | times | one of --times, --every, --at | Times of day, comma-separated (09:00,17:30); with --weekdays (mon,wed,fri, ranges like mon-fri, sat-sun or fri-mon across the weekend, or the API's numbers 0 Sunday to 6 Saturday, comma-separated; default every day) and --tz (an IANA name; default the workspace's time zone) |
--every | duration | Every N minutes or hours (30m, 6h; 5 minutes to a week) | |
--at | time | One run at this time (same formats as posts create --publish-at) | |
--channel | string | Yes | The channels the chosen posts go to, repeatable (a post's own targets win) |
--pick | string | No | next (default: the oldest waiting draft), all (every waiting draft) or selected (the --post ids; --post alone implies it) |
--source | string | No | When nothing waits: none (default), agent (ask your agent; it shows up in staqly work) or endpoint (--endpoint-url, signed with --endpoint-secret; generated when empty and then printed once by create). --endpoint-url alone implies endpoint |
--disabled, --enable, --disable | boolean | No | Create paused; turn on or off on update |
Needs schedules:manage. See Schedules.
The queue: jobs
staqly jobs list [--status queued|running|done|failed|cancelled] [--kind publish_post|publish_target|target_timeout|run_schedule|refresh_channel] [--limit 50]
staqly jobs stats # queued, running, failed, done in the last 24 hours, by kind
staqly jobs get job_123
staqly jobs retry job_123 # a failed job, now (posts:publish)
staqly jobs cancel job_123 # a queued jobSee Jobs.
The media library: media
staqly media list [-q name] [--kind image|video|audio] [--used used|unused] [--folder mfd_...|root] [--tag hero] [--sort size_desc] [--limit 60 --offset 60]
staqly media upload cover.png # prints the media ID
staqly media update med_oalirbz67y4m4cqxq5dw --alt "A runner at the finish line" [--name cover.png]
staqly media usage med_oalirbz67y4m4cqxq5dw # the posts that use it
staqly media get med_oalirbz67y4m4cqxq5dw -o cover.png
staqly media get med_oalirbz67y4m4cqxq5dw --width 320 -o thumb.jpg
staqly media get med_oalirbz67y4m4cqxq5dw | file - # piped: bytes go to stdout
staqly media delete med_oalirbz67y4m4cqxq5dw # to Deleted for 30 days (--permanent: for good)
staqly media restore med_oalirbz67y4m4cqxq5dw # back out of Deleted
staqly media list --deleted # what is in Deleted; --duplicates: identical files
staqly media keep med_oalirbz67y4m4cqxq5dw # keep it, its identical copies go to Deleted
staqly media mv med_oalirbz67y4m4cqxq5dw med_t4w6pz2kqm7xr3nb5vya --folder mfd_k2s7pbk7igv6 # or --folder root
staqly media tag med_oalirbz67y4m4cqxq5dw --add hero,summer-sale [--remove draft]
staqly media folders list # the tree with file counts, then the tags
staqly media folders create "Clients" [--in mfd_k2s7pbk7igv6]
staqly media folders rename mfd_k2s7pbk7igv6 "Clients 2027"
staqly media folders move mfd_k2s7pbk7igv6 --to root
staqly media folders delete mfd_k2s7pbk7igv6 # its files and folders move up; no file is deletedAttach an uploaded file to a post with --media on posts create or posts update.
| Flag | Type | Required | Description |
|---|---|---|---|
--output, -o | path | No | get: file to write; - writes to stdout. Without it the bytes go to stdout when it is piped, otherwise to a new file named after the original in the current directory |
--width | integer | No | get: thumbnail width for images, 160, 320, 640 or 1280 |
-q, --kind, --used, --sort, --limit, --offset | No | list: filter by name, type (image, video, audio) and use (used, unused), sort (created_desc, created_asc, size_desc, name_asc) and page | |
--name, --alt | string | No | update: new file name, alt text |
The CLI never writes binary data to a terminal and never overwrites a file it named itself.
Organize collections: collections
staqly collections list [--archived]
staqly collections create "Product updates" [--description <text>] [--icon <emoji or name>] [--color "#3b82f6"] [--processing-timeout 2h]
staqly collections get "Product updates"
staqly collections rename "Product updates" "Updates"
staqly collections timeout Updates 2h # your own automation's targets fail when nothing is reported within 2 hours; `default` goes back to the server's
staqly collections archive Updates [--undo]
staqly collections delete Updates [--yes] # needs collections:delete; without --yes: type the name to confirmdelete permanently removes the collection with its posts and schedules.
| Flag | Type | Required | Description |
|---|---|---|---|
--description | string | No | create: description shown under the name |
--icon | string | No | create: an emoji or icon name |
--color | string | No | create: colour of the collection's dot and calendar bar, e.g. #3b82f6. The dashboard offers green #22c55e, blue #3b82f6, violet #8b5cf6, pink #ec4899, red #ef4444, orange #f97316, yellow #eab308 and teal #14b8a6 |
--processing-timeout | duration | No | create: how long a target on your own automation may wait for a result, e.g. 30m or 2h (1 minute to 24 hours); without it the server default applies. collections timeout <collection> <duration or default> changes it later |
--archived | boolean | No | list: include archived collections |
--undo | boolean | No | archive: unarchive instead |
--yes | boolean | No | delete: skip the confirmation; required when stdin is not a terminal |
Follow the activity log: events
staqly events tail # last 20 events, oldest first
staqly events tail --follow --type 'post.*' # poll for new events (--interval 2s)
staqly events tail -f --json | jq -c 'select(.type=="post.failed")'
staqly events tail --post post_1 --collection Updates --actor-type api_key --limit 100
staqly events tail --channel ch_1 --schedule sch_1 --from 2026-10-01 --to 2026-10-14--follow tracks the highest seq printed and pages back so no event is missed. An exact
--type is filtered by the server; a glob such as post.* by the CLI, after the server returns
the last --limit events, so a glob can print fewer than --limit lines.
| Flag | Type | Required | Description |
|---|---|---|---|
--follow, -f | boolean | No | Keep polling for new events |
--interval | duration | No | Poll interval with --follow (default 2s, minimum 500ms) |
--limit | integer | No | How many recent events to print first (default 20, max 200) |
--type | string | No | Exact event type such as post.published, or a glob such as post.* |
--actor-type | string | No | local, session, api_key or system |
--actor-id | string | No | Only events by this actor, e.g. an API key ID such as key_r7y65khtqa23kj4btvur |
--from | time | No | Events created at or after this time, RFC3339 or YYYY-MM-DD |
--to | time | No | Events created before this time, RFC3339 or YYYY-MM-DD |
--collection | string | No | Collection name or ID |
--post, --channel, --schedule | string | No | Only events about this post, channel or schedule |
Export the log
staqly events export > events.ndjson
staqly events export --format csv --from 2026-10-01 --to 2026-11-01 -o october.csv
staqly events export --type 'post.*' --collection "Product updates" | jq -c 'select(.type=="post.failed")'export streams every matching event, oldest first, with no page size. CSV columns are
created_at,id,type,actor_type,actor_id,actor_name,collection_id,post_id,data.
Credential and session audit events (api_key.*, session.*, recovery_key.*, access_token.*, sign_in_method.*, device*) are
included only for admin principals.
| Flag | Type | Required | Description |
|---|---|---|---|
--format | string | No | ndjson (default, one JSON event per line) or csv |
--from, --to | time | No | Created at or after / before this time, RFC3339 or YYYY-MM-DD |
--type | string | No | Exact event type, or a prefix ending in * such as post.* (sent to the server as type_prefix=post.; can't be combined with --type-prefix) |
--type-prefix | string | No | Literal type prefix, e.g. post. |
--actor-type, --actor-id | string | No | As in tail |
--collection, --post | string | No | Collection name or ID; post ID |
--output, -o | path | No | Write to this file instead of stdout |
Notify other systems: webhooks
staqly webhooks list # secrets masked
staqly webhooks create --url https://example.com/hook --events post.published,post.failed [--collection Updates] [--description <text>] [--timeout-ms 10000] [--max-attempts 8] [--inactive]
staqly webhooks get wh_123 # secret masked
staqly webhooks update wh_123 [--url <url>] [--events 'post.*'] [--collection Updates | --all-collections] [--description <text>] [--timeout-ms 10000] [--max-attempts 8] [--enable | --disable] [--rotate-secret]
staqly webhooks test wh_123
staqly webhooks deliveries [--webhook wh_123] [--status pending|succeeded|failed] [--limit 50]
staqly webhooks retry dlv_123
staqly webhooks delete wh_123create prints the signing secret once. --events defaults to * (all events); an unknown
event name or glob (a typo such as post.publised) is refused with 422. See
Webhooks.
update changes only the fields you pass:
| Flag | Type | Required | Description |
|---|---|---|---|
--url | string | No | New receiver URL (with this command, set the API address with --api-url) |
--events | string list | No | Replaces the event names or globs, comma-separated or repeated |
--collection | string list | No | Replaces the collection filter (names or IDs, repeatable) |
--all-collections | boolean | No | Deliver events of every collection again |
--description | string | No | New description; --description "" clears it |
--timeout-ms | integer | No | Delivery timeout, 1000 to 60000 ms |
--max-attempts | integer | No | Delivery attempts, 1 to 20 |
--enable, --disable | boolean | No | Activate or deactivate the webhook |
--rotate-secret | boolean | No | Issue a new signing secret. It is printed once, and the old secret stops verifying deliveries |
Tell people: alerts
staqly alerts list
staqly alerts get nch_oalirbz67y4m4cqxq5dw # settings, secrets masked
staqly alerts create --type ntfy --name Phone --config topic=staqly-alerts-7f3k
echo '{"webhook_url":"https://hooks.slack.com/services/T000/B000/XXXX"}' | staqly alerts create --type slack --name Team --config-json -
staqly alerts create --type email --name Ops --config host=smtp.example.com --config port=587 --config from=staqly@example.com --config to=ops@example.com --events 'post.*' --digest 10m
staqly alerts update nch_oalirbz67y4m4cqxq5dw [--name <text>] [--config key=value] [--config-json <json>] [--events <patterns>] [--collection C | --all-collections] [--digest 5m] [--enable | --disable]
staqly alerts test nch_oalirbz67y4m4cqxq5dw # exit 1 if the message failed
staqly alerts deliveries nch_oalirbz67y4m4cqxq5dw [--limit 50]
staqly alerts delete nch_oalirbz67y4m4cqxq5dw| Flag | Type | Required | Description |
|---|---|---|---|
--type | string | Yes | create only: telegram, slack, discord, ntfy or email; it can't change later |
--name | string | Yes | Alert name (required for create, optional for update) |
--config | key=value | No | One setting, repeatable. port and priority are numbers, to a comma list; key=null removes a setting on update |
--config-json | JSON | No | Settings as a JSON object; - reads stdin and @path a file, which keeps secrets out of your shell history. --config values are merged on top |
--events | string list | No | Event patterns; default post.published, post.failed, channel.needs_reconnect. Unknown patterns are refused with 422 |
--collection | string list | No | Only these collections (names or IDs, repeatable) |
--all-collections | boolean | No | update only: tell about every collection again |
--digest | duration | No | Batch events within this window into one message, e.g. 5m; 0 sends immediately |
--disabled | boolean | No | create only: create the alert disabled |
--enable, --disable | boolean | No | update only: enabling resets the failure counter |
--limit | integer | No | deliveries only: how many recent messages to show (default 50, max 200) |
create needs at least one setting. Required settings per type: Telegram bot_token and
chat_id; Slack webhook_url, or bot_token and channel; Discord the same as Slack; ntfy
topic; email host, from and to. On update, settings merge into the stored ones. Secrets
are stored encrypted and always shown masked. Managing alerts needs the alerts:manage scope.
See Alerts.
Manage API keys: keys
staqly keys list
staqly keys create --name "Content agent" --scope collections:read,posts:read,posts:write,media:write [--expires 2027-01-01] # printed once; expires in a year unless --expires is sooner
staqly keys rotate key_123 # revoke + new key with the same name and scopes, printed once
staqly keys revoke key_123API keys are a cloud-mode feature: the Community Edition has none (create answers 409), its agents use
the access token. Managing keys needs the admin scope. create and rotate print the new key once; store it right away.
Manage the workspace: workspace, plans
staqly workspace get
staqly workspace rename "Acme content" # admin
staqly workspace timezone Europe/Warsaw # admin; repeats and schedules without a zone follow it
staqly workspace usage # plan, usage vs limits, trial end, suspension
staqly workspace export start [--wait] [--output staqly-export.zip] [--timeout 30m]
staqly workspace export status exp_oalirbz67y4m4cqxq5dw
staqly workspace export download exp_oalirbz67y4m4cqxq5dw [-o backup.zip]
staqly plans list # cloud plans, monthly and yearly prices, limitsThe export is a ZIP with collections, posts, channels (without credentials), schedules, events,
media files and webhooks (secrets redacted). Only one export runs at a time, and finished
archives are deleted after 24 hours. Exports need the admin scope.
| Flag | Type | Required | Description |
|---|---|---|---|
--wait | boolean | No | export start: poll until the export is ready; exits 1 if it failed and 3 on timeout |
--output, -o | path | No | export start: download the finished ZIP to this file (implies --wait). export download: file to write, - for stdout; without it the ZIP goes to stdout when piped, else to the server's file name in the current directory |
--timeout | duration | No | export start: how long --wait waits (default 30m) |
--interval | duration | No | export start: poll interval (default 2s) |
Deleting the workspace and rotating the recovery key are not CLI commands: the API only allows them from a signed-in dashboard session, not with an API key or the local principal.
Pay for a plan: billing
Cloud workspaces with Stripe billing enabled (admin key):
staqly billing status # plan, subscription status, paid-through date, suspended
staqly billing checkout --plan plus --yearly # Stripe Checkout link; a paying workspace is switched instead
staqly billing portal # prints a link to change the card, get invoices or cancelThe links are for a person to open in a browser; the plan changes when Stripe confirms the payment.
Sign in other browsers: connect, sessions
staqly connect # one-time sign-in link, valid 5 minutes (Community Edition)
staqly sessions list # signed-in browsers (admin)
staqly sessions revoke ses_oalirbz67y4m4cqxq5dwconnect prints a link that signs one browser in once, for example a phone on your network. The
link is a credential: open it only on your own device. The server issues these links only to the
dashboard or the Community Edition's access token, so a cloud server answers 403 to an API key. When the link
points at localhost, the CLI warns that other devices can't open it; set PUBLIC_DASHBOARD_URL on the
server.
Manage profiles: login, logout, profiles
staqly login --url https://app.staqly.ai --key stk_... --profile work [--no-verify]
staqly profiles
staqly profiles use work
staqly logout --profile workConnect an agent: mcp config
Prints a ready-to-use MCP configuration for an agent client, pointing at the current profile's
server (<url>/mcp):
staqly mcp config claude # `claude mcp add` command + .mcp.json snippet
staqly mcp config codex # `codex mcp add` command + ~/.codex/config.toml
staqly mcp config opencode --stdio # opencode.json for docker exec -i staqly mcp-stdio
staqly mcp config cursor --snippet > .cursor/mcp.json
staqly mcp config windsurf # ~/.codeium/windsurf/mcp_config.json snippetHTTP snippets reference $STAQLY_API_KEY (an API key, or the Community Edition's access token:
export STAQLY_API_KEY=$(staqly serve token)); the key itself is never printed. --stdio against
the Community Edition needs no key. --snippet prints only the config file snippet, and --container changes the
container name used with --stdio (default staqly).
Other commands
staqly version [--json]
staqly completion bash|zsh|fish
staqly --help
staqly <command> --help