CLI

pipelines

Manage cloud pipelines, runs, groups and reviews

Editframe enables pipelines for each organization after you tell us about your footage. Until then, every pipeline request returns 404.

Manage pipelines and pipeline runs in Editframe's cloud. Unlike the other commands, pipelines calls the cloud API directly: it needs no local runtime, and it does not use --project, --bridge or the other runtime options.

bash

export EDITFRAME_TOKEN="$EDITFRAME_API_KEY"
npx editframe pipelines create speech-check --definition speech-check.json
npx editframe pipelines runs create PIPELINE_ID FILE_ID --watch

Requests authenticate with the API key in EDITFRAME_TOKEN. The host is --host, then EDITFRAME_HOST, then https://editframe.com. editframe pipelines --help prints the full usage.

Commands

Command	Description
`list [--include-archived] [--paused]`	List pipelines. `--paused` lists only paused pipelines.
`get PIPELINE_ID`	Show a pipeline and its latest version
`create NAME --definition FILE.json`	Create a pipeline with version 1
`archive PIPELINE_ID`	Archive a pipeline
`pause PIPELINE_ID`	Pause every version of a pipeline. No more runs or steps start. Steps in progress finish, and new runs wait until you resume the pipeline.
`resume PIPELINE_ID`	Resume a paused pipeline. The waiting runs and steps start.
`versions list PIPELINE_ID`	List versions
`versions get PIPELINE_ID VERSION`	Show one version
`versions create PIPELINE_ID --definition FILE.json`	Create a version, or find the version with the same definition
`endpoints list [--include-archived]`	List endpoints
`endpoints get ENDPOINT_ID`	Show an endpoint
`endpoints create URL [--header "Name: value"]... [--timeout-ms N] [--callback-timeout-ms N] [--max-concurrency N]`	Register an endpoint. The signing secret is printed once.
`endpoints update ENDPOINT_ID [--timeout-ms N] [--callback-timeout-ms N] [--max-concurrency N]`	Change the named limits. Requests sent after the change use them.
`endpoints archive ENDPOINT_ID`	Archive an endpoint
`runs create PIPELINE_ID FILE_ID [--version N] [--group GROUP_ID] [--external-id ID] [--metadata KEY=VALUE]... [--idempotency-key KEY] [--watch]`	Start a run. `--group` joins an open group, which must use the same version.
`runs create PIPELINE_ID --source-url URL [options]`	Start a run on a video at an HTTPS URL, read in place instead of uploaded. It takes the options of `runs create` with a file ID. See URL sources.
`runs list [--status S] [--pipeline-id ID] [--file-id ID] [--group-id ID] [--metadata KEY=VALUE]...`	List runs, newest first
`runs get RUN_ID [--watch]`	Show a run and its steps
`runs cancel RUN_ID`	Cancel a run
`runs retry RUN_ID [--watch]`	Retry a failed or cancelled run
`runs outputs RUN_ID NAMESPACE`	Show the outputs a run wrote to a namespace
`groups create PIPELINE_ID [FILE_ID...] [--files IDS.txt] [--source-url URL]... [--version N] [--close] [--max-in-flight N] [--external-id ID] [--metadata KEY=VALUE]... [--idempotency-key KEY]`	Create a group, with one run for each file and each `--source-url`. `--close` accepts no more runs, so the group completes when these runs finish.
`groups close GROUP_ID`	Accept no more runs. The group completes when every run finished.
`groups update GROUP_ID --max-in-flight N`	Change how many runs of the group run at the same time (1–100,000). When you raise it, waiting runs start immediately.
`groups list [--status open\|closed\|completed] [--pipeline-id ID] [--metadata KEY=VALUE]...`	List groups
`groups get GROUP_ID`	Show a group and its run counts
`groups report GROUP_ID`	Show run counts, queue wait, run duration, input minutes, failures by step and recorded usage
`reviews list [--status pending\|approved\|rejected\|expired] [--run-id ID] [--pipeline-id ID]`	List reviews, oldest first
`reviews get REVIEW_ID`	Show a review
`reviews approve REVIEW_ID [--patch FILE.json]`	Approve a review, with optional corrections
`reviews reject REVIEW_ID [--reason TEXT]`	Reject a review
`upload FILE [--run PIPELINE_ID] [--version N] [--group GROUP_ID] [--watch] [--resume FILE_ID] [--external-id ID] [--metadata KEY=VALUE]... [--fragment-duration SECONDS] [--concurrency N]`	Upload a video as verified fragments, and optionally start a run on it
`metadata FILE_ID [NAMESPACE]`	List the namespaces pipelines wrote on a file, or show one

Options

Option	Description
`--json`	Print API responses as JSON
`--host URL`	API host (default: `EDITFRAME_HOST` or `https://editframe.com`)
`--limit N`	Page size for lists and outputs (1–100)
`--cursor CURSOR`	Continue a list from the next cursor of a previous page
`--watch`	Poll the run until it finishes. Exits with status 1 when the run fails or is cancelled.
`--metadata KEY=VALUE`	Add an entry to the `metadata` of a run or group that you create, or list only the runs or groups that have the entry. Repeat it for more entries. Values can contain `=`.

Notes

A --files file has one file ID per line. Text after # is ignored.
A --patch file is { "annotations": [...], "reviewed_kinds": [...] }. See Approve.
upload --run starts the run as soon as the file exists, so segment steps process fragments as they arrive. Its default idempotency key is derived from the file, the pipeline and the version, so --resume finds the same run. When an upload stops, the CLI prints the command that resumes it.