MCP App Store
by KeenEthics
ViddyScribe icon

ViddyScribe MCP Server for ChatGPT

By ViddyScribe Inc.

Overview

ViddyScribe creates audio descriptions, captions and transcripts for videos, so people who are blind or have low vision can follow what happens on screen. Connect your ViddyScribe account to browse your library, read and edit description scripts, change the narration voice, translate scripts, check your team's credits, and start new audio descriptions from a video link or a video already in your library. Anything that uses credits, such as creating a description, rendering a video, or translating a script, always shows the cost first and only starts after you confirm. Finished videos, audio, scripts, transcripts, and captions can be downloaded from the links the app returns. Video files can't be uploaded in chat: use a link, or upload the file in ViddyScribe first. Requires a ViddyScribe account on app.viddyscribe.com.

How to connect

  1. ChatGPT: Open the ViddyScribe listing in the ChatGPT app directory and choose Connect. ChatGPT listing ViddyScribe exposes 19 tools here.

Tools

check_credits

ChatGPT
The team's credit balance, usage this cycle, and pending-job headroom.

create_description

ChatGPT
Start a NEW audio-description job for a video already in the library, using the video_id from list_library or a prior quote - e.g. an extended_ad version of a video that only has standard_ad, or a new language. Without confirm: returns the credit quote only. With confirm=true and max_credits set to the approved quote_credits: starts the job (spends credits). If the price is higher, nothing starts and a new quote is returned.

create_from_url

ChatGPT
Create an audio description from a video URL (YouTube, Vimeo, or a direct video link). First call WITHOUT confirm: ingests the URL and returns a credit quote plus a video_id - show the quote to the user. To actually start the job (spends credits), call again with confirm=true, that video_id, and max_credits set to the approved quote_credits; if the price is higher, nothing starts and a new quote is returned. ad_type: standard_ad fits narration into natural pauses; extended_ad pauses the video for longer descriptions. voice: a value from list_voices (omit for the default voice).

export_captions

ChatGPT
Caption/subtitle file for a version in vtt or srt (defaults: current version, vtt). For reading what is said, use get_transcript instead.

get_cues

ChatGPT
The audio-description script as structured cues (cue_key, start/end seconds, text) for editing. Use before save_edited_script. Defaults to the current version.

get_downloads

ChatGPT
Short-lived (~5 min) download URLs for a version's FINISHED outputs (described video, AD audio) for the user to download. Not for processing. Defaults to the current version; empty until render_video has completed.

get_script

ChatGPT
The audio-description script text for a version (defaults to the current version). Good for reviewing or quoting in chat.

get_status

ChatGPT
Status of one project (generation_id from list_library or create_from_url): per-version progress, ETA, and which artifacts are ready to download.

get_transcript

ChatGPT
Timestamped transcript of a version: what people say in the video, with the audio descriptions woven in. Use it for context about the video before writing or editing descriptions. Defaults to the current version.

ingest_status

ChatGPT
Progress of a URL fetch started by create_from_url (by tracking_id).

list_languages

ChatGPT
Every language ViddyScribe can describe and translate into, as exact codes (e.g. en-US, es-ES, cmn-TW) with names. Use these codes for the language arguments of the other tools.

list_library

ChatGPT
Browse or search the team's audio-description projects. Returns newest first; pass cursor from a previous page for more.

list_voices

ChatGPT
Narration voices available for a language (a code from list_languages, or a prefix like es). Pass a voice's value to create_from_url, create_description, or set_voice.

render_video

ChatGPT
Produces the described video: ViddyScribe voices any unvoiced cues, mixes, and renders the version server-side. Without confirm: returns the credit quote only. With confirm=true and max_credits set to the approved quote: starts the render. If the price is above max_credits, nothing starts and a new quote is returned. If the version's last render failed, this retries it once for free instead. Track it with get_status.

save_edited_script

ChatGPT
Save changes to the AD script as a NEW version (free, never overwrites the original). Use it whenever the user wants descriptions shortened, rewritten, retimed, removed, or added: pass only the cues that change, keyed by cue_key from get_cues; unchanged cues are kept. Returns the new version_id to pass to render_video, which is how the edited script becomes a video.

set_voice

ChatGPT
Change the narration voice of an existing version (free; nothing is voiced yet). Then call render_video: its quote includes any new speech in this voice. Defaults to the current version.

translate_script

ChatGPT
Translate a finished AD script into another language (exact code from list_languages, e.g. es-ES) as a new language version of the same video. Costs 0.1 credits per video minute (minimum 0.1), charged only if it succeeds. Without confirm: returns the credit quote only. With confirm=true and max_credits set to the approved quote: starts it. If the price is above max_credits, nothing starts and a new quote is returned. Defaults to the current version.

wait_for_completion

ChatGPT
Wait (server-side, up to max_wait_seconds, default 90) for a generation or render to finish and return the outcome: ready (with download links when a video was rendered), failed (with the reason), or still_running (with an ETA). Call once right after confirming a job. Defaults to the current version.

whoami

ChatGPT
The signed-in ViddyScribe user, their team(s), role, and plan.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/viddyscribe"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/viddyscribe)