Videosays · Video to Text

Transcribe video links and extract spoken scripts or subtitles from Douyin, Xiaohongshu, Bilibili, YouTube, TikTok, and other supported platforms. Use for single or batch transcription, timestamped transcripts, SRT/VTT export, credit balance, and transcription history.

wegofuture@wegofuture

Install

openclaw skills install @wegofuture/videosays

Videosays Video to Text

Use npx -y videosays@latest to submit video links and retrieve transcript text or subtitles. Commands resolve the current npm latest release at execution time, so a CLI update does not require republishing this Skill. The CLI sends the configured API key and submitted links/share text to the configured Videosays API. CLI source: https://github.com/xwchris/videosays-agent-tools/tree/main/packages/cli .

Accept video links or copied share text when the user asks for a spoken script, transcript, or subtitle file. The CLI does not accept local video files. Source accessibility and supported content depend on the service response.

Language and API

Reply in the user's language: Chinese for Chinese requests, English for English requests. Explain CLI errors in that language; preserve command names, task IDs, and subtitle timestamps.
Return the actual transcript in its original language. Summarize or translate it only when requested, and distinguish that work from the source transcript.
The CLI's default API is https://api.videosays.cn, serving the same backend and accounts as .com. Language does not select a different backend. Honor an explicitly configured VIDEOSAYS_API_URL.
Existing global installations can update with npm install -g videosays@latest. Skill commands use npx -y videosays@latest, which resolves the current published CLI without relying on a globally installed copy. Honor an explicit VIDEOSAYS_API_URL and never switch domains automatically.

Requirements

Node.js 18 or newer
npx

Authentication

Before the first transcription in a session, run:

bash

npx -y videosays@latest whoami

If authentication is missing, run:

bash

npx -y videosays@latest 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 -y videosays@latest login --api-key "$VIDEOSAYS_API_KEY". Never print or reveal the API key.

Single Link

Submit one link:

bash

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

Submission returns quickly. It normally prints:

text

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:

bash

npx -y videosays@latest 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:

bash

npx -y videosays@latest status "<task-id>" --format timeline
npx -y videosays@latest status "<task-id>" --format srt
npx -y videosays@latest 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.

The CLI automatically retries a transient network disconnect with the same submission ID while rotating resolved API addresses. It does not retry an explicit HTTP error and never changes domains. If the command still ends before printing a Task ID, do not resubmit with a new id. Retry with the same --submission-id when it is available; otherwise check npx -y videosays@latest 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.

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.
Submit once:

bash

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

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

bash

npx -y videosays@latest batch status "<batch-id>"

Repeat status checks until the batch reaches completed, partial, failed, cancelled, or paused. Stop polling when paused and explain what is needed to continue.

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 is available; otherwise check recent tasks and report the ambiguous outcome if it cannot be recovered.

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, continuation.required, or stopReason: insufficient_credits, completed Tasks are preserved and unstarted Tasks may be skipped. Ask the user to top up, then after confirmation run:

bash

npx -y videosays@latest 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:

bash

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

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

Other Commands

bash

npx -y videosays@latest balance
npx -y videosays@latest history
npx -y videosays@latest doctor
npx -y videosays@latest 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.

text

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.

For connection errors, run npx -y videosays@latest doctor and share its output with support. The command probes the configured API's resolved addresses without sending the API key.