Videosays

Videosays Video Transcription

Use npx videosays to submit video links and retrieve transcript text or subtitles. The CLI sends the configured API key and submitted links/share text to Videosays.

Requirements

• Node.js 18 or newer
• npx

Authentication

Before the first transcription in a session, run:

npx videosays whoami

If authentication is missing, run:

npx videosays login

Ask the user to open the printed authorization URL, sign in, and approve the CLI. If the user explicitly provides an API key, run npx videosays login --api-key "$VIDEOSAYS_API_KEY". Never print or reveal the API key.

Single Link

Submit one link:

VIDEOSAYS_CLIENT_SURFACE=agent_skill VIDEOSAYS_CLIENT_NAME=videosays-skill npx videosays transcribe "<video-link-or-share-text>"

Submission returns quickly. It normally prints:

VIDEOSAYS_TASK_PENDING
task_id=<task-id>
status=<status>
next=videosays status <task-id>

Capture task_id. Wait a reasonable interval, then run the printed one-shot status command:

npx videosays status "<task-id>"

If the task is still running, status immediately prints its current state and the next command. Repeat status checks until the command prints transcript content or a stable error. Preserve the requested format on status checks:

npx videosays status "<task-id>" --format timeline
npx videosays status "<task-id>" --format srt
npx videosays status "<task-id>" --format vtt

Each submission gets a client-generated submissionId sent as Idempotency-Key. If a command ends ambiguously, repeat the same command with the same --submission-id <uuid>; the server returns the original Task safely. Equivalent active work for the same account is reused, and completed work is reused by default. Use --force-new only when a fresh transcription is intentional. Treat transcribe only as creation: capture its returned server Task ID and use status for every later check. Never resubmit a link as a status check with a new submission id.

If submission ends with a network error or timeout before printing a Task ID, do not automatically resubmit with a new id. Retry with the same --submission-id when it is available; otherwise check npx videosays history for a recent matching task and report the ambiguous outcome if it cannot be recovered.

Multiple Links

When the user provides two or more links, use one server batch. Never build a shell loop, use xargs, start parallel transcribe commands, or submit the links individually.

1. Write one link or share text per line to a temporary text file. Duplicate lines keep their batch positions, while their effective status/result follows the canonical Task by default. Use --force-new when the user explicitly requests duplicate processing.
2. Submit once:

VIDEOSAYS_CLIENT_SURFACE=agent_skill VIDEOSAYS_CLIENT_NAME=videosays-skill npx videosays batch links.txt

3. Capture the server-generated batchId from stdout.
4. Wait a reasonable interval, then make a one-shot status request:

npx videosays batch status "<batch-id>"

5. Repeat status checks until the batch reaches completed, partial, failed, or cancelled, or reports a resumable paused state.

While a batch is running, batch status uses a lightweight status response. When the batch finishes, the CLI retrieves the complete results once. Do not replace this with per-Task polling.

Every batch <file> submission sends a client-generated Idempotency-Key. Repeat an ambiguous submission with the same --submission-id <uuid> to receive the original server Batch ID; an equivalent active batch for the same account also converges on one Batch ID. Completed work is reused by default. Use --force-new only for an intentional fresh batch. Batch submission and status commands return promptly. Do not rerun the input file as a status check, do not invent a Batch ID, and do not use batch resume.

If batch submission ends before printing a Batch ID, do not automatically submit the file with a new id: the server may already have accepted it. Retry with the same --submission-id if one was printed, or report the ambiguous outcome and get confirmation before creating a replacement batch with --force-new.

Videosays creates every batch item as an ordinary Task and runs those Tasks through the normal queue. Each Task must reserve credit atomically before provider submission, so the balance cannot be overspent. If the batch reports paused or stopReason is insufficient_credits, completed Tasks are preserved and unstarted Tasks wait to be resumed. Ask the user to top up, then after confirmation run:

npx videosays batch continue "<batch-id>"

Then continue using batch status with the same Batch ID.

Optional Interactive Waiting

Only use --wait when a human explicitly wants the terminal to remain attached:

npx videosays transcribe "<video-link>" --wait
npx videosays batch links.txt --wait

Agents must use the default immediate-return workflow so every tool call produces prompt, structured stdout.

Other Commands

npx videosays balance
npx videosays history
npx videosays batch cancel "<batch-id>"

Errors

Read stderr when a command exits non-zero. Do not treat error output or a pending receipt as transcript content.

Error: <message>
Code: <error-code>
Next: <recommended-command>
Recharge: <billing-url>

For insufficient_credits, do not repeatedly resubmit. Report the balance issue and recharge URL. For media or link errors such as media_resolve_failed, media_unavailable, or media_inaccessible, ask for another accessible video link.

Links

• Website: https://videosays.com/?utm_source=videosays_skill&utm_medium=agent_skill&utm_campaign=videosays_agent_skill
• API docs: https://videosays.com/docs?utm_source=videosays_skill&utm_medium=agent_skill&utm_campaign=videosays_agent_skill&utm_content=api_docs
• CLI: https://www.npmjs.com/package/videosays