openclip-transcription
Transcribe video or audio to JSON, SRT, and VTT with speaker diarization and time codes, using OpenClip. FREE with just an OpenClip account, no subscription. Use when asked to "transcribe this", "get me a transcript", "make subtitles", "generate an SRT", "who said what in this recording", "pull quotes from this video", or "what was said at 4:30".
How do I install this agent skill?
npx skills add https://github.com/openclip-app/agent-skills --skill openclip-transcriptionIs this agent skill safe to install?
- Gen Agent Trust Hubpass
This skill provides instructions for using the OpenClip platform to transcribe audio and video files. It describes standard OAuth authentication and remote MCP server setup using the official OpenClip domain. No security issues were detected.
- Socketpass
No alerts
- Snykpass
Risk: LOW · No issues
What does this agent skill do?
OpenClip: transcription
transcribe turns a video or audio file into a time-coded transcript. It is FREE, it needs an
OpenClip account but no subscription and no credits. Output formats: JSON (structured, with
timings), SRT, and VTT.
Setup (once)
OpenClip is a remote MCP server at https://openclip.app/mcp. Sign in with your OpenClip
account when the OAuth prompt appears, there is no API key to copy.
- Claude Code:
claude mcp add --transport http openclip https://openclip.app/mcp, then/mcpto authorize. - Claude Desktop / web: Settings, Connectors, "Add custom connector", paste the URL, sign in.
- Cursor (
.cursor/mcp.json):{ "mcpServers": { "openclip": { "url": "https://openclip.app/mcp" } } } - Header-only clients: mint an MCP token at
openclip.app/settings/connectand connect tohttps://openclip.app/mcp/keywithAuthorization: Bearer <token>.
Sanity check: call get_account.
The loop
- Upload.
create_upload(filename, content_type)returns anidand anupload_url. PUT the raw bytes toupload_url, then pass theidasvideo. Do NOT callcomplete_upload, that starts the paid clipping pipeline. Free tools read the uploaded file directly. Videos already inlist_videoswork as inputs too. - Transcribe.
transcribe(video, language?, diarize?).language- omit to auto-detect. Set it when you already know the language and want to stop a short or noisy clip being misdetected.diarize- defaults totrue, labels who is speaking. Turn it off for single-speaker audio to keep the output clean.
- Poll. Returns a
tool_jobhashid. Pollget_tool_job_status(tool_job)every 5 to 10 seconds.queued/processingare in flight,completedreturns permanent CDNoutputsURLs (one per format: json, srt, vtt),failedreturnserror. Most jobs finish inside a minute or two. - Read the output. Fetch the JSON output for programmatic work (segments with start and end times, speaker labels). Hand the user the SRT or VTT URL when they want a subtitle file.
Two different transcript tools
Do not mix these up:
transcribe- the FREE tool. Works on any file you upload. This skill.get_transcript- reads the transcript of a video that already went through the PAID clipping pipeline. Supportsstart_ms/end_mswindows at sentence or word level. Use it when you already have a clippingjob_id, not for a fresh upload.
Rules
- Always poll. Transcription is async, it never returns the text synchronously.
- Times are milliseconds.
- Output URLs are permanent CDN links, not signed or expiring.
- Free usage is rate limited per day with a per-file size cap. The error text explains how to lift a limit.
Example prompts to actions
- "Transcribe this interview" -
create_upload, PUT bytes,transcribe(video), poll, return the transcript. - "I need an SRT for this video" - same, then hand back the
srtoutput URL. - "Who said what in this meeting recording?" -
transcribe(video, diarize=true), poll, read the JSON output and group segments by speaker. - "Pull the best quotes out of this talk" - transcribe, poll, read the JSON, then select quotes with their time codes.
- "It is in Dutch, transcribe it" -
transcribe(video, language="nl"), poll.
Related
To then cut the video at those time codes, see openclip-video-editing (edit_video with
operation="trim"). To find the moments worth cutting automatically, see openclip-clipping.
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/openclip-app/agent-skills/openclip-transcription">View openclip-transcription on skillZs</a>