yt-subtitles-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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": true
} |
| logging | {} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_video_contextA | Start here for video summaries, key takeaways, comparisons or Q&A. Return metadata, chapters and a timestamped transcript page in one request. With language omitted, select the original audio language when available, otherwise authored captions before automatic captions. Partial coverage and unavailable transcripts are explicit. This tool supplies evidence; the client agent writes the summary. It cannot analyze visual-only content. |
| get_video_infoA | Get a video's title, channel, duration, upload date and views as untrusted data. |
| list_transcript_languagesB | List exact available subtitle codes, distinguishing authored and automatic captions. |
| get_video_transcriptA | Continue reading timestamped source text for summarization or questions about a video. Use the language returned by get_video_context when continuing its transcript. With no language specified, prefer the original audio language, then authored captions. Offset and limit count characters. Repeat with next_offset until it is null. No translation or visual analysis is performed. Use list_transcript_languages for exact codes. |
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 reasonably distinct primary role: context bootstraps, info gives lightweight metadata, list_transcript_languages enumerates codes, and get_video_transcript paginates. However, get_video_info's metadata overlaps with get_video_context's returned metadata, and context also returns transcript, creating mild ambiguity about when to use info vs. the composite tool.
All names are snake_case with a verb-first pattern (get_/list_), which is predictable. There is a minor resource-naming inconsistency: three tools scope to 'video' while list_transcript_languages scopes to 'transcript', so the resource vocabulary isn't fully uniform.
Four tools is on the thin side but appropriate for a focused transcript/metadata server where each tool has a clear jobs-to-be-done. The only questionable inclusion is get_video_info, which largely duplicates a subset of get_video_context.
The surface covers metadata, chapters, language enumeration, transcript retrieval with pagination, and explicit handling of partial/unavailable coverage, which is solid lifecycle coverage for this domain. Gaps like in-transcript search or translation are either out of scope or explicitly excluded by the descriptions.