# vid2context (npm · vid2context-mcp)

Turn any video into context for AI agents: transcripts, metadata, chapters, timestamps.

- Trust score: 75/100 (medium)
- Change this week: +18
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- remote · `vid2context.com`: 36/100, [markdown](https://verifymcp.io/servers/nirholas-vid2context/vid2context.md), [page](https://verifymcp.io/servers/nirholas-vid2context/vid2context)
- npm · `vid2context-mcp`: 75/100 (this document), [markdown](https://verifymcp.io/servers/nirholas-vid2context/vid2context-mcp.md), [page](https://verifymcp.io/servers/nirholas-vid2context/vid2context-mcp)

## Channel facts

- Registry: `npm`
- Package: `vid2context-mcp`
- Version: `0.2.1`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-21.

- **Supply Chain Security**: 100/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - No install/post-install scripts declared.
  - No production dependencies, so there is no dependency health to assess.
- **Provenance & Transparency**: 48/100
  - Source repository is publicly reachable at the declared URL.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (Apache-2.0).
  - Actively maintained (last published 1 days ago).
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 76/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 7217 tokens (~189/item across 38 items; 37 tools + 1 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 23/100
  - Stability observed for 7 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
  - Structured output schemas are declared (16% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 37 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 39 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.
  - Supports UI / widget rendering.

## Install

### How do I install the vid2context MCP server?

vid2context runs locally as an npm package, launched with npx -y vid2context-mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add nirholas-vid2context -- npx -y vid2context-mcp
```

### Cursor

```json
{
  "mcpServers": {
    "nirholas-vid2context": {
      "command": "npx",
      "args": [
        "-y",
        "vid2context-mcp"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "nirholas-vid2context": {
      "command": "npx",
      "args": [
        "-y",
        "vid2context-mcp"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add nirholas-vid2context -- npx -y vid2context-mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "nirholas-vid2context": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "vid2context-mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add nirholas-vid2context --command npx --arg -y --arg vid2context-mcp
```

### Hermes

```yaml
mcp_servers:
  nirholas-vid2context:
    command: "npx"
    args: ["-y", "vid2context-mcp"]
```

### Netclaw

```json
{
  "McpServers": {
    "nirholas-vid2context": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "vid2context-mcp"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add nirholas-vid2context -t stdio -c npx -a -y vid2context-mcp
```

### Other

```json
{
  "mcpServers": {
    "nirholas-vid2context": {
      "command": "npx",
      "args": [
        "-y",
        "vid2context-mcp"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-21 (score 75, +1)

No change was recorded against any check on this day. Stability & Change Management went from 20 to 23. That category is still filling its 30-day observation window: 6 days of observed history at the previous scan, 7 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-19 (score 74, +2)

- [security regression] Tool safety: pass → unverified
- [security regression] Known CVEs: pass → unverified
- [security] Stability: Stability not yet verified: we do not have a sandbox capture of the MCP schema this version of the package serves yet.
- [functional regression] Tool coverage: 21% → 16%
- [functional regression] Schema quality: 151 → 189
- [functional regression] Capabilities: pass → unverified
- [functional regression] Tool coverage: 100 → unverified
- [functional regression] Schema quality: 100 → unverified
- [functional improvement] Stability: unverified → 0.17
- [functional] Package version: 0.1.1 → 0.2.1
- [functional] Package version: 0.1.1 → 0.2.0

### 2026-11-15 (score 72, +15)

- [security improvement] Malware scan: unverified → pass

### 2026-11-14 (score 57)

First indexed and scored.

## MCP tools (37)

### `get_transcript` (~664 tokens)

Get video transcript

The full transcript of one video or local media file, with a header (title, channel, URL, duration, how the transcript was made, chapters). With timestamps (the default) every line starts with [m:ss] so you can cite moments. Uses creator captions, then platform auto-captions, then speech-to-text. Cost: speech is about 12k tokens per hour of video. For videos longer than about 20 minutes, call get_outline first and read only the sections you need with get_section, or set maxTokens to receive the transcript in pages (each page ends with a cursor for the next one). Formats: text (Plain transcript text. With timestamps, one "[m:ss] line" per caption.); json (The full VideoContext as JSON: metadata, chapters, timed segments, and how it was produced.); markdown (Markdown transcript: title and metadata, chapters as headings, paragraphs prefixed with timestamp links. Options: linkTimestamps, includeComments, includeVisual.); srt (SubRip (.srt) captions: at most 2 lines of 42 characters per cue, no overlaps, for players and editors.); vtt (WebVTT (.vtt) captions for the HTML <track> element, with speaker voice tags when known.); pack (Context pack: one Markdown file for one or many videos, written for an AI model to read, with citation instructions. Options: maxTokens, includeComments, includeVisual, linkTimestamps.); ass (Advanced SubStation Alpha (.ass) subtitles with a readable default style, for mpv, VLC, ffmpeg burn-in and Aegisub. Bilingual subtitles show the second language in a smaller style.); ttml (TTML / IMSC (.ttml) timed text XML for broadcast, streaming packagers and players that read TTML or DFXP.); sbv (YouTube SubViewer (.sbv) captions, the format YouTube Studio imports and exports.); lrc (LRC (.lrc) timed lyrics, one [mm:ss.xx] line per cue, for music players and karaoke apps.); csv (CSV (.csv) with one row per segment: start,end,speaker,text in seconds, plus secondary_text for bilingual output. For spreadsheets and data tools.)

Input parameters:

- `cursor` (string): Continuation cursor from a previous partial result (like "seg:412"). Use the same other arguments
- `format` (string): Output format (see list_formats). Default: text
- `language` (string): Preferred transcript language (BCP-47, e.g. en, de). Default: the spoken language
- `maxTokens` (integer): Largest result to return, in estimated tokens. Longer transcripts come back in pages with a continuation cursor
- `mode` (string): Transcript strategy: auto (captions, then speech-to-text), captions only, stt (always transcribe), video (multimodal model). Default: auto
- `timestamps` (boolean): Prefix lines with [m:ss] timestamps. Default: true
- `url` (string, required): Video URL (YouTube and more), a YouTube video id, or a local media file path

### `get_metadata` (~75 tokens)

Get video metadata

Title, channel, description, duration, publish date, chapters, tags and language for a video URL or local file, without fetching the transcript. Cheap: a few hundred tokens, mostly the description.

Input parameters:

- `url` (string, required): Video URL (YouTube and more), a YouTube video id, or a local media file path

### `list_formats` (~65 tokens)

List output formats

The output formats get_transcript and get_section accept (text, json, markdown, srt, vtt, pack, ...), with what each is for. combinesVideos marks formats that can merge several videos into one document (export_context_pack). Cheap: under 500 tokens.

Output parameters:

- `formats` (array)

### `get_outline` (~198 tokens)

Get video outline

Start here for any video longer than about 20 minutes. Returns the metadata and the video's chapters (or sections of about five minutes when it has none), each with its start time, a timestamp link and its estimated token count, plus the total. Then read only what you need with get_section (by chapter index or title, or by time). Fetches the transcript once (saved for later calls); the outline itself costs about 300 to 1,500 tokens.

Input parameters:

- `language` (string): Preferred transcript language (BCP-47, e.g. en, de). Default: the spoken language
- `mode` (string): Transcript strategy: auto (captions, then speech-to-text), captions only, stt (always transcribe), video (multimodal model). Default: auto
- `url` (string, required): Video URL (YouTube and more), a YouTube video id, or a local media file path

Output parameters:

- `chapters` (array)
- `durationSec` (number)
- `metadata` (object)
- `next` (string)
- `sections` (string): chapters: the video's own chapters; time: sections of about five minutes
- `title` (string)
- `totalTokens` (number)
- `transcript` (object)
- `url` (string)

### `get_section` (~296 tokens)

Get a transcript section

One chapter or time range of a video's transcript, with the same header as get_transcript. Pass chapter (an index or title from get_outline) or from/to (seconds, m:ss or h:mm:ss); from alone reads to the end, to alone reads from the start. Cost: the section's token count from get_outline (about 1k tokens per 5 minutes of speech).

Input parameters:

- `chapter`: A chapter index from get_outline, or a chapter title (exact, prefix, then substring match, case-insensitive)
- `format` (string): Output format. Default: text
- `from`: Start time: seconds (754), m:ss (12:34) or h:mm:ss (1:02:03). Default: the start
- `language` (string): Preferred transcript language (BCP-47, e.g. en, de). Default: the spoken language
- `mode` (string): Transcript strategy: auto (captions, then speech-to-text), captions only, stt (always transcribe), video (multimodal model). Default: auto
- `timestamps` (boolean): Prefix lines with [m:ss] timestamps. Default: true
- `to`: End time (exclusive), same forms as from. Default: the end
- `url` (string, required): Video URL (YouTube and more), a YouTube video id, or a local media file path

### `get_comments` (~106 tokens)

Get video comments

Top comments of a video, with creator replies and pinned comments flagged. Useful for corrections, follow-up questions and audience consensus that the video itself lacks. Cost: about 50 tokens per comment; limit defaults to 50. Only sources whose provider can read comments support this.

Input parameters:

- `limit` (integer): Most comments to return. Default: 50
- `url` (string, required): Video URL (YouTube and more), a YouTube video id, or a local media file path

Output parameters:

- `comments` (array)
- `count` (number)
- `title` (string)
- `url` (string)

### `list_items` (~163 tokens)

List the videos in a playlist, channel or feed

Expand a playlist, channel, podcast feed or page into its videos (a single video URL returns itself). A video URL that also names a playlist (watch?v=...&list=...) returns the video plus `list`, the playlist; pass playlist: true to expand the playlist instead. Use it to see what a list contains before reading, or pass the list straight to add_to_collection. Cost: about 40 tokens per item; limit defaults to 50.

Input parameters:

- `limit` (integer): Most items to return. Default: 50
- `playlist` (boolean): When url is a video link that also names a playlist (list=), list the playlist instead of the one video
- `url` (string, required): Playlist, channel, feed or page URL

Output parameters:

- `count` (number)
- `input` (string)
- `items` (array)
- `list` (object)

### `export_context_pack` (~168 tokens)

Export a context pack

One Markdown document for many videos (the pack format): a header per video with title, channel, URL and chapters, then each transcript in timestamped paragraphs. Set maxTokens to fit a budget: the pack trims transcripts to fit and says what it left out. Cost: about 12k tokens per hour of video unless maxTokens caps it. Prefer get_outline, get_section and search_transcript when you only need parts.

Input parameters:

- `collection` (string): Export this saved collection (id or name)
- `limit` (integer): Most videos taken from each playlist, channel or feed. Default: 50
- `maxTokens` (integer): Token budget for the whole pack
- `urls` (array): Video, playlist, channel or feed URLs (lists are expanded into their videos)

### `list_models` (~123 tokens)

List models

The models (rungs) this server can use for speech-to-text, LLM, embedding, video and visual work, in free-first order: tier (local, free, keyless, paid), quality 1 to 5, capabilities (json, tools, vision, stream, ...), context window, whether it is enabled, and the names of the env vars that would enable it. Never shows key values. Cost: about 80 tokens per model; filter with kind.

Input parameters:

- `kind` (string): Only this kind of model. Default: all kinds

Output parameters:

- `defaultPolicy` (string)
- `models` (array)

### `search_transcript` (~200 tokens)

Search transcripts

Find where something is said across saved video transcripts, instead of reading whole transcripts (about 12k tokens per hour). Hybrid search: keyword matching for exact names and model numbers, plus meaning-based matching for paraphrases (when an embedding model is available). Returns the best time ranges with title, [m:ss], highlighted text and a timestamp link. Syntax: "exact phrase" must appear, -word excludes. Scope with collection (a collection id) or urls (videos are fetched and saved first if needed).

Input parameters:

- `collection` (string): Only search this collection (its id)
- `limit` (integer): Hits to return. Default 10
- `query` (string, required): What to look for. "quoted phrases" are required, -word excludes a word
- `rerank` (boolean): Rerank the top results with a free LLM. Slower; default false
- `urls` (array): Only search these videos

### `build_brief` (~205 tokens)

Build a cited brief

Build a cited brief across videos on the server with a free LLM (map-reduce for large collections). Every claim cites a video, timestamp and quote that were checked against the transcript; points are marked consensus (two or more videos agree), disputed (both sides cited) or single-source, and claims that failed the check are listed separately, uncited. If you can read the transcripts yourself, the `brief` prompt is cheaper and usually better; use this for collections too large for your context.

Input parameters:

- `collection` (string): A saved collection id
- `format` (string): markdown (default): the brief as Markdown; json: the full Brief object
- `question` (string): What the brief should answer, e.g. "Which G-Wagon year should I buy?"
- `template` (string): brief (default) or buyer-guide, tutorial-steps, comparison, summary
- `urls` (array): Up to 25 video URLs or ids

### `extract` (~171 tokens)

Extract structured data

Extract structured JSON from videos with a template (buyer-guide, tutorial-steps, comparison, summary, brief) or your own JSON Schema. The answer is validated against the schema (one repair attempt), and every citation ({videoId, startSec, quote}) is checked against the transcript: verified ones gain a title and timestamp url, failed ones are removed and listed in "unverified".

Input parameters:

- `collection` (string): A saved collection id
- `question` (string): Focus for the extraction
- `schema` (object): A JSON Schema (type, properties, required, items, enum, min/max limits). The result is {data, citations}.
- `template` (string): Template id; default summary when no schema is given
- `urls` (array): Up to 25 video URLs or ids

### `list_collections` (~38 tokens)

List collections

List research collections with how many videos and sources each has. A collection groups videos on one topic and can follow playlists, channels and podcast feeds.

### `create_collection` (~94 tokens)

Create collection

Create a research collection. sources are playlists, channels or podcast feeds the collection follows: add_to_collection or a refresh ingests their videos, and watch lists pick up new uploads. Returns the collection with its id.

Input parameters:

- `description` (string): What the collection is about
- `name` (string, required): Collection name, for example "G-Wagon"
- `sources` (array): Playlist, channel or feed URLs to follow

### `add_to_collection` (~107 tokens)

Add to collection

Ingest videos into a collection now: each URL is a video or a list (playlist, channel, feed), and lists are expanded. Transcripts come from captions first, speech-to-text only when needed. Videos already in the collection are skipped and per-video failures are reported without stopping. Returns the updated collection and a summary (added, skipped, failed with hints).

Input parameters:

- `collection` (string, required): Collection id
- `urls` (array, required): Video, playlist, channel or feed URLs

### `get_collection` (~69 tokens)

Get collection

A collection with a summary of each video (id, title, channel, duration, URL, transcript origin), the sources it follows, inputs waiting for the next refresh and its watch list. Read transcripts with get_transcript, get_outline or search_transcript.

Input parameters:

- `collection` (string, required): Collection id

### `ask` (~116 tokens)

Ask videos

Answer a question from the transcripts of one or more videos, or a saved collection, with server-side retrieval and the free-first LLM chain. Every factual sentence cites a numbered source; each source is a timestamp link to the moment that supports it. Says so plainly when the videos do not cover the question. Use it when you want a cited answer without reading whole transcripts.

Input parameters:

- `collection` (string): A collection id
- `question` (string, required): The question
- `urls` (array): Video URLs or ids like youtube:abc

### `research_videos` (~214 tokens)

Research videos

Search 1 to 10 videos (or a playlist, capped at 10) for a question and get only the relevant excerpts: ranked by BM25, each with the surrounding context, [m:ss] and a timestamp link, under a token cap. Use it instead of reading whole transcripts (about 12k tokens per hour of speech). Put exact phrases in double quotes to boost them. Lexical: if nothing matches, it lists words the videos do use so you can retry with a synonym.

Input parameters:

- `contextSec` (number): Seconds of context before and after each match. Default 20
- `format` (string): text (compact Markdown, default) or json
- `maxExcerpts` (integer): Excerpts per video. Default 3
- `maxTokens` (integer): Cap on the whole result. Default 2000
- `query` (string, required): Words to look for; "quoted phrases" are boosted
- `urls` (array, required): Video URLs, or one playlist URL

### `list_caption_languages` (~63 tokens)

List caption languages

List the caption tracks a video has (language, name, creator or auto-generated) and, for YouTube, the languages YouTube can machine-translate them into. Call it before get_transcript to pick a language.

Input parameters:

- `url` (string, required): Video URL

### `get_key_moments` (~88 tokens)

Get key moments

The moments of a video worth jumping to, picked by the free-first LLM chain from the timestamped transcript. Every timestamp is checked against the real transcript (snapped within 5 seconds, dropped otherwise); the result says how many were dropped and which rung answered.

Input parameters:

- `count` (integer): How many moments. Default 8
- `url` (string, required): Video URL

### `get_skip_segments` (~229 tokens)

Get sponsor and skip segments

Community-marked SponsorBlock segments of a YouTube video: sponsor reads, self-promotion, subscribe reminders, intros, outros, previews and filler, each with its [m:ss] range, a timestamp URL and how much of the transcript it covers (lines, word share and the first words). Use it to leave ad copy out of summaries and briefs: never cite text inside these segments as a claim about the topic. Default categories: sponsor, selfpromo, interaction, intro, outro, preview, filler. All: sponsor, selfpromo, interaction, intro, outro, preview, hook, filler, music_offtopic, poi_highlight, exclusive_access, chapter. poi_highlight comes back as `highlight` (a point, not a skip); a whole-video label comes back as `fullVideo`. Other platforms return no segments. Results carry the SponsorBlock attribution (CC BY-NC-SA 4.0).

Input parameters:

- `categories` (array): Categories to look up. Default: the env default or the seven above
- `url` (string, required): YouTube video URL or id

### `translate_transcript` (~312 tokens)

Translate transcript

A video's transcript in another language, with every segment and timestamp kept. Free first: the transcript as is when it is already in that language, then the platform's own track or machine translation (YouTube translated captions), then LLM translation on the free-first rung chain. The source transcript is saved unchanged; the translation is an added layer. Text output starts with a header naming the video and how it was translated; other formats come with a second content block reporting the path used and the attempts. Costs about as many tokens as the transcript itself (about 12k per hour of speech).

Input parameters:

- `format` (string): Output format: text, json, markdown, srt, vtt, pack, ass, ttml, sbv, lrc, csv. Default text. json returns the full result, source transcript included
- `language` (string, required): Target language, BCP-47 (en, de, pt-BR)
- `llm` (boolean): Allow LLM translation when the platform has no track in the language. Default true; false uses only the platform's tracks
- `mode` (string): How to get the source transcript: auto (captions, then speech-to-text), captions, stt or video. Default auto
- `timestamps` (boolean): Prefix lines with [m:ss] where the format allows. Default false
- `url` (string, required): Video URL (YouTube and more), a YouTube video id, or a local media file path

### `export_obsidian` (~180 tokens)

Export to an Obsidian vault

Write one Markdown note per video into a folder of an Obsidian vault on this machine: frontmatter properties, chapters as headings, paragraphs prefixed with [m:ss] timestamp links, plus an index note with [[wikilinks]] when a collection is exported. Notes vid2context wrote before are updated; other existing notes are never replaced unless overwrite is true.

Input parameters:

- `collection` (string): A saved collection id to export
- `folder` (string): Folder inside the vault. Default: the collection name, or "vid2context"
- `overwrite` (boolean): Replace existing notes not written by vid2context for the same video. Default false
- `urls` (array): Video, playlist, channel or feed URLs to export
- `vaultPath` (string, required): Absolute path of the vault (or any folder) to write into

### `export_notion` (~128 tokens)

Export to Notion

Create one Notion page per video under a parent page (chapters as headings, paragraphs with timestamp links, a bookmark to the video). Re-exporting updates the same pages instead of duplicating them. Needs NOTION_TOKEN; the parent page comes from parentPageId or NOTION_PARENT_PAGE_ID.

Input parameters:

- `collection` (string): A saved collection id to export
- `parentPageId` (string): Notion page id or URL to create the pages under. Default: NOTION_PARENT_PAGE_ID
- `urls` (array): Video, playlist, channel or feed URLs to export

### `find_duplicates` (~143 tokens)

Find reuploads and clips

Compare the transcripts of several videos and group reuploads (the same video uploaded again) and clips (an excerpt of a longer video, with the [m:ss] where it starts in the original and the range it covers). Use it before writing a brief across many videos: count each group once (representativeOf maps every video to its original), or a clip agreeing with its own source looks like consensus. Related pairs share passages without being copies. Works on transcripts only; re-dubbed or translated reuploads are not detected.

Input parameters:

- `collection` (string): A saved collection id
- `urls` (array): Video URLs or ids; playlists and channels are expanded

### `watch_live` (~129 tokens)

Watch a live stream

Transcribe a live stream (a live YouTube video or an HLS .m3u8 URL) while it runs. New lines arrive about every 30 seconds as progress notifications; the result is the full transcript with [stream time | UTC] stamps when the stream ends or maxMinutes pass. Default 10 minutes, at most 60. For recordings use get_transcript.

Input parameters:

- `maxMinutes` (number): Minutes to follow the stream. Default 10, capped at 60
- `url` (string, required): Live YouTube URL or HLS (.m3u8) URL

### `open_transcript_viewer` (~128 tokens)

Open transcript viewer

Show a video transcript in an interactive viewer (chapters, search, clickable [m:ss] timestamps, one-click citations, send selected lines to the chat) in clients that support MCP Apps, such as ChatGPT and Claude. The text result starts with a header (title, channel, URL, chapters); clients without MCP Apps get the full timestamped transcript as text.

Input parameters:

- `language` (string): Preferred transcript language (BCP-47, e.g. en, de)
- `url` (string, required): Video URL (YouTube and more) or a local file path

Output parameters:

- `chapters` (array)
- `metadata` (object)
- `segments` (array): Transcript lines as [startSec, endSec, text]
- `timestampBase` (string|null): Link that opens the video at a moment: replace {t} with whole seconds. Null when there is no deep link.

### `get_caption_options` (~103 tokens)

List caption tracks for download

Every caption track of one video (YouTube, any site yt-dlp reads, or a local file with sidecar or embedded subtitles): language, name, creator or automatic, which provider lists it, the languages the platform can machine-translate into, and whether speech-to-text can make a transcript when there are none. Nothing is downloaded. Call it before download_captions.

Input parameters:

- `url` (string, required): Video URL, YouTube id or local file path

### `download_captions` (~490 tokens)

Download captions or subtitles

Save the captions of a video as srt, vtt, ass, ttml, sbv, lrc, csv, txt, json, markdown or the original file the site serves, to a folder on this machine; returns the path. offsetSec retimes the cues, maxLineChars and maxLines reflow them, and secondary adds a second language under each caption. Works for every source vid2context reads, through its transcript pipeline (no yt-dlp needed), with yt-dlp as a second path; for a video without captions, mode auto (default) transcribes the audio with speech-to-text. languages: one or several codes, or ["all"] for every track; several files come as one .zip. Files go to dir, else VID2CONTEXT_DOWNLOAD_DIR, else ~/Downloads/vid2context.

Input parameters:

- `allowAuto` (boolean): Accept the platform's automatic captions. Default true
- `allowTranslated` (boolean): Accept the platform's machine translation. Default true
- `dir` (string): Folder to save to: absolute, ~/..., or relative to the default folder
- `format` (string): srt, vtt, ass, ttml, sbv, lrc, csv, txt, json, markdown, original (default srt; original keeps the site's file)
- `languages` (array): Language codes, e.g. ["en","de"], or ["all"]. Default: the spoken language
- `maxLineChars` (integer): Longest caption line in characters (cue formats, default 42)
- `maxLines` (integer): Most lines per cue (default 2)
- `mode` (string): auto (captions, else speech-to-text; default), captions (never transcribe), stt (always transcribe)
- `offsetSec` (number): Shift every cue by this many seconds (negative: earlier); timed formats only
- `secondary` (string): A second language shown under each caption, e.g. "de" (bilingual formats)
- `timestamps` (boolean): txt only: one "[m:ss] line" per caption
- `url` (string, required): Video URL, YouTube id or local file path
- `via` (string): auto (default), pipeline or ytdlp
- `zip` (boolean): Return a .zip even for one file

### `grab_frames` (~228 tokens)

Grab still frames from a video

Save still images (png default, or jpg) of a video at given times to a folder on this machine. One time gives one image, several a .zip. Web videos are downloaded first (only the needed part when the times are close together). Files go to dir, else VID2CONTEXT_DOWNLOAD_DIR, else ~/Downloads/vid2context. Download only media you have the rights to: your own uploads, public-domain or Creative Commons works, or videos the site lets you download.

Input parameters:

- `dir` (string): Folder to save to: absolute, ~/..., or relative to the default folder
- `format` (string): png (default) or jpg
- `quality` (string): Download size for web videos (smaller side), default 1080
- `times` (array, required): Times: seconds, m:ss or h:mm:ss
- `url` (string, required): Video URL, YouTube id or local file path
- `width` (integer): Width in pixels; default the source size
- `zip` (boolean): Return a .zip even for one frame

### `export_clip` (~249 tokens)

Export an animated GIF or WebP

Save a short part of a video as an animated GIF (default) or WebP to a folder on this machine. from and to are seconds or m:ss (default 5 s from from); fps default 10 and width default 480, both capped by the server limits (VID2CONTEXT_MAX_CLIP_*). For a video clip with sound use download_media with from and to. Download only media you have the rights to: your own uploads, public-domain or Creative Commons works, or videos the site lets you download.

Input parameters:

- `dir` (string): Folder to save to: absolute, ~/..., or relative to the default folder
- `format` (string): gif (default) or webp
- `fps` (integer): Frames per second, default 10
- `from` (string|number): Start: seconds, m:ss or h:mm:ss. Default 0
- `loop` (boolean): Repeat forever (default true)
- `to` (string|number): End. Default from + 5 s
- `url` (string, required): Video URL, YouTube id or local file path
- `width` (integer): Width in pixels, default 480

### `download_batch` (~337 tokens)

Download many videos, audio or captions

Save several videos at once: a list of URLs, playlists or channels (expanded up to limit), each as video, audio or captions only. Failed items are reported and do not stop the rest. zip: true returns one .zip with every file. Video formats: mp4, webm, mkv; audio: mp3, m4a, opus, flac, wav; captions: srt, vtt, ass, ttml, sbv, lrc, csv, txt, json, markdown, original. Download only media you have the rights to: your own uploads, public-domain or Creative Commons works, or videos the site lets you download.

Input parameters:

- `dir` (string): Folder to save to: absolute, ~/..., or relative to the default folder
- `format` (string): Format for that kind, e.g. mp4, mp3 or srt
- `kind` (string): What each video becomes. Default video
- `languages` (array): captions: language codes or ["all"]
- `limit` (integer): Most videos in total. Default and cap: VID2CONTEXT_MAX_PLAYLIST_ITEMS
- `mode` (string): captions: auto, captions or stt
- `playlist` (boolean): Expand video links that also name a playlist (watch?v=...&list=...)
- `quality` (string): Video: 1080, 720, ...; mp3: kbps
- `urls` (array, required): Video, playlist or channel URLs, YouTube ids or file paths
- `zip` (boolean): Return one .zip with every saved file

### `get_media_options` (~110 tokens)

Get video and audio download options

Qualities, formats and estimated file sizes for downloading one video (YouTube or any site yt-dlp supports, or a local file): video sizes per quality for mp4, webm and mkv, audio sizes for mp3 (by bitrate), m4a, opus, flac and wav, plus chapters and subtitle languages. Nothing is downloaded. Call it before download_media to pick a quality that fits.

Input parameters:

- `url` (string, required): Video URL, YouTube id or local file path

### `download_media` (~632 tokens)

Download a video or its audio

Save a video as mp4, webm or mkv, or its audio as mp3, m4a, opus, flac or wav, to a folder on this machine; returns the file paths. Works for YouTube and the other sites yt-dlp supports, and converts local files. Defaults: video mp4 up to 1080p (H.264 + stereo AAC, plays everywhere), audio mp3 at the best VBR quality, with title, cover art and chapters embedded. Keep a part with from/to (seconds or m:ss) or chapter (index or title); cut sponsor reads with removeSegments (YouTube). A playlist or channel URL downloads its first `limit` videos. Files go to dir, else VID2CONTEXT_DOWNLOAD_DIR, else ~/Downloads/vid2context. Download only media you have the rights to: your own uploads, public-domain or Creative Commons works, or videos the site lets you download.

Input parameters:

- `burnSubtitles` (string): Burn one language's subtitles into the picture, e.g. "en" (video only; re-encodes, so slower)
- `chapter` (string): Keep one chapter: 0-based index or title (see get_media_options)
- `dir` (string): Folder to save to: absolute, ~/..., or relative to the default folder
- `embedChapters` (boolean): Chapter markers. Default true
- `embedMetadata` (boolean): Title, channel and date tags. Default true
- `embedThumbnail` (boolean): Cover art. Default true
- `exactCuts` (boolean): Re-encode at the cut points so a trimmed video starts exactly at from (slower)
- `format` (string): mp4 (default for video), webm, mkv, mp3 (default for audio), m4a, opus, flac, wav
- `from` (string): Start of the part to keep: seconds, m:ss or h:mm:ss
- `kind` (string): video or audio. Default: from format, else video
- `limit` (integer): Playlists and channels: how many videos to download. Default 10
- `quality` (string): Video: best, 2160, 1440, 1080, 720, 480, 360, 240, 144 (smaller side in pixels; default 1080). Audio: best, 320, 256, 192, 160, 128, 96, 64 (mp3 kbps; default best)
- `removeSegments` (array): SponsorBlock categories to cut out (YouTube)
- `splitChapters` (boolean): Also save one file per chapter, in a folder next to the full file. Not with from/to/chapter
- `subtitles` (array): Subtitle languages to embed in a video, e.g. ["en"]
- `to` (string): End of the part to keep
- `url` (string, required): Video, playlist or channel URL, YouTube id, or local file path

### `download_thumbnail` (~144 tokens)

Save a video thumbnail

Save a video's thumbnail (cover image) as jpg, png or webp to a folder on this machine; returns the path. For YouTube and other sites it takes the largest image the site lists; for a local file, its cover art, else a frame from the first tenth of the video. Files go to dir, else VID2CONTEXT_DOWNLOAD_DIR, else ~/Downloads/vid2context.

Input parameters:

- `dir` (string): Folder to save to: absolute, ~/..., or relative to the default folder
- `format` (string): jpg (default), png or webp
- `url` (string, required): Video URL, YouTube id or local file path

### `digest_videos` (~194 tokens)

Digest videos with your own model

Build one cited digest of several videos (a playlist, a collection, or a list of URLs) using MCP sampling: the server asks the client's model to take notes on each part of each transcript, then combine them. Every claim cites [Title @ m:ss](link), and every timestamp is checked against the real transcript. No server key is needed. If this client cannot sample, the result is a chunk plan with the exact instructions, so you can do the same work yourself.

Input parameters:

- `collection` (string): A collection id or name; its saved videos are digested
- `maxTokensPerVideo` (integer): Most transcript tokens sent in one sampling call; longer videos are split by chapter or time. Default 6000
- `question` (string): A question the digest should answer
- `urls` (array): Video URLs, playlists or channels (expanded into their videos), saved video ids or titles

### `choose_caption_track` (~78 tokens)

Choose a caption track

List a video's caption tracks (languages, creator or auto-generated) and ask the user to pick one, then return that transcript with [m:ss] timestamps. Without a picker in this client, or when the user declines, returns the track list so you can call get_transcript with a language.

Input parameters:

- `url` (string, required): Video URL

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/nirholas-vid2context/vid2context-mcp#diagnostics

## Score history

- 2026-11-21: 75
- 2026-11-20: 74
- 2026-11-19: 74
- 2026-11-18: 72
- 2026-11-17: 72
- 2026-11-16: 72
- 2026-11-15: 72
- 2026-11-14: 57

## Common questions

### What is the vid2context MCP server?

vid2context is an MCP server listed in the public MCP registry as io.github.nirholas/vid2context. Turn any video into context for AI agents: transcripts, metadata, chapters, timestamps. This page covers its npm package (vid2context-mcp).

### Is the vid2context MCP server safe to use?

vid2context scores 75 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 21 November 2026. It declares no install or post-install scripts. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the vid2context MCP server expose?

vid2context exposes 37 tools: get_transcript, get_metadata, list_formats, get_outline, get_section, and 32 more. Their descriptions and schemas cost roughly 7,037 tokens of context every time the server is loaded.

### Is the vid2context MCP server still maintained?

vid2context is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the vid2context MCP server under?

vid2context declares the Apache-2.0 licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

### What did VerifyMCP test on the vid2context MCP server?

VerifyMCP ran 17 checks on vid2context, covering Malware scan, Known CVEs, Install scripts, Dependency health and 13 more. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- npm package: https://www.npmjs.com/package/vid2context-mcp
- Socket report: https://socket.dev/npm/package/vid2context-mcp
- Repository: https://github.com/nirholas/vid2context
- Website: https://vid2context.com/
- Changelog RSS feed: https://verifymcp.io/servers/nirholas-vid2context/vid2context-mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/nirholas-vid2context/vid2context-mcp/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/nirholas-vid2context/vid2context-mcp.json
- HTML version of this page: https://verifymcp.io/servers/nirholas-vid2context/vid2context-mcp
