Video Transcript MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HF_ENDPOINT | No | Set to https://hf-mirror.com for China network optimization | |
| HF_HUB_DISABLE_XET | No | Disable Xet storage (avoids download errors) | 1 |
| TRANSCRIPT_OUTPUT_DIR | No | Output directory for transcripts | ~/.video-transcript-mcp/output |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| transcribe_urlA | Transcribe a video from URL using three-tier strategy. Tier 1: Extract platform subtitles (zero cost, fastest) Tier 2: Download audio + Whisper local transcription (offline, free) Tier 3: Return guidance for closed platforms (Xiaohongshu/WeChat) For videos under 30 minutes, use sync mode (default) for direct results. For longer videos, set async_mode=true to get a task_id, then poll with get_transcript_status. Adjust segment_minutes for 1h+ videos. |
| transcribe_fileA | Transcribe a local audio/video file using Whisper. Supports mp3, m4a, mp4, wav, flac, and other FFmpeg-compatible formats. For files over 30 minutes, use async_mode=true. Adjust segment_minutes for 1h+ files to reduce segment count. |
| get_transcript_statusA | Check the status of an async transcription task. Returns the task status, progress, and result (if completed). Poll this periodically until status is 'completed' or 'failed'. |
| list_transcriptsA | List all completed transcripts. Returns a list of completed transcription tasks with metadata. Use get_transcript_status with a task_id to get full transcript text. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 4 tools
Each tool has a clearly distinct purpose: transcribe_file processes local media, transcribe_url handles remote videos, get_transcript_status polls async tasks, and list_transcripts shows completed transcripts. The descriptions explicitly differentiate these, leaving no ambiguity.
All four tools follow a consistent verb_noun pattern with snake_case (transcribe_file, transcribe_url, get_transcript_status, list_transcripts). The verbs clearly indicate the action and the nouns the resource, forming a predictable and coherent naming scheme.
Four tools is well-scoped for a video transcription server, covering the essential operations (transcribe local/remote, check status, list results) without unnecessary overlap or bloat. This is an appropriate size for the domain.
The tool set covers the full transcription workflow: synchronous and async transcription, status polling, and listing completed results. Minor gaps like cancellation or direct retrieval of a transcript by ID are missing, but these are not critical for the core use case.