Augent

The audio & video layer for agents. 22 local MCP tools. No cloud, no API keys.

Security Scan

The skill's runtime instructions, required binaries, and optional environment variables are coherent with its stated purpose of local audio/video processing — no unexplained credentials or network exfiltration are requested.

Purpose & Capability

The name/description match the requested binaries and tools: augent-mcp, ffmpeg, yt-dlp, and aria2c are expected for downloading and processing media; the tool set (transcription, diarization, demucs, sentence-transformers, etc.) is consistent with the stated features.

Instruction Scope

SKILL.md instructs the agent to run local MCP tools, save files to user directories, and optionally read a Twitter/X auth token for Spaces recording. There are no instructions to read unrelated system secrets or transmit data to unexpected external endpoints. It does specify writing to user Download/Desktop and to an Obsidian vault (for graph rebuild), which is within the tool's scope but worth noting.

Install Mechanism

Install uses package managers (uv and a pip alternative appears in SKILL.md metadata). This is a standard approach, but the pip/uv install will pull many heavy ML dependencies (faster-whisper, sentence-transformers, pyannote, demucs). The registry listing shows uv as primary; SKILL.md metadata also mentions pip — minor inconsistency but explainable as alternate install methods.

Credentials

No required credentials. One optional variable (AUGENT_AUTH_TOKEN) is explicitly for recording X/Twitter Spaces and is reasonable. Other env vars are directory paths for downloads, notes, and memory. No unrelated secrets or broad credential requests are present.

Persistence & Privilege

always: false (normal). The skill will create/read files in user-specified/default directories (~Downloads, ~/Desktop, ~/.augent/memory) and may interact with an Obsidian vault for graph rebuilding — this is consistent with its note-taking/transcription features and does not indicate elevated platform privileges or edits to other skills.

Assessment

This skill appears to do what it says: local media download, transcription, speaker ID, clipping, and local TTS. Before installing: 1) Confirm you trust the Augent package source (GitHub link) because installation pulls many ML dependencies. 2) Be aware the tool writes media and transcripts to your Downloads, Desktop, and ~/.augent/memory (you can change these paths). 3) The optional AUGENT_AUTH_TOKEN is only needed if you want to record X/Twitter Spaces — do not provide other credentials. 4) If you use an Obsidian vault, note the skill can read/write to it when rebuilding graphs; back it up if concerned. 5) If you want tighter control, inspect the package (uv/pip install target) before installing or run it in an isolated environment/container.

Augent — Audio & Video Intelligence for AI Agents

Augent is an MCP server that gives your agent 22 tools for audio and video intelligence. Download from 1000+ sites via yt-dlp and aria2c, transcribe in 99 languages via faster-whisper, search by keyword or meaning via sentence-transformers, take notes, identify speakers via pyannote-audio, detect chapters, separate audio via Demucs v4, export clips, extract visual frames, record X/Twitter Spaces (requires user-configured auth token in ~/.augent/auth.json), and generate speech via Kokoro TTS. All processing runs locally. Downloads are saved to ~/Downloads/, notes and clips to ~/Desktop/, transcription memory to ~/.augent/memory/.

Config

{
  "mcpServers": {
    "augent": {
      "command": "augent-mcp"
    }
  }
}

If augent-mcp is not in PATH, use python3 -m augent.mcp as the command instead.

Install

Install via the ClawHub install button above, or use uv tool install augent for the base package or uv tool install "augent[all]" for all features. FFmpeg is required for audio processing.

Tools

Augent exposes 22 MCP tools:

Core

Tool	Description
download_audio	Download audio from video URLs at maximum speed. Supports YouTube, Vimeo, TikTok, Twitter/X, SoundCloud, and 1000+ sites. Uses aria2c multi-connection + concurrent fragments.
transcribe_audio	Full transcription of any audio file with per-segment timestamps. Returns text, language, duration, and segments. Cached by file hash.
search_audio	Search audio for keywords. Returns timestamped matches with context snippets. Supports clip export.
deep_search	Semantic search — find moments by meaning, not just keywords. Uses sentence-transformers embeddings.
search_memory	Search across ALL stored transcriptions in one query. Keyword or semantic mode.
take_notes	All-in-one: download audio from URL, transcribe, and save formatted notes. Supports 5 styles: tldr, notes, highlight, eye-candy, quiz.
clip_export	Export a video clip from any URL for a specific time range. Downloads only the requested segment.

Analysis

Tool	Description
chapters	Auto-detect topic chapters with timestamps using embedding similarity.
search_proximity	Find where two keywords appear near each other (e.g., "startup" within 30 words of "funding").
identify_speakers	Speaker diarization — identify who speaks when. No API keys required.
separate_audio	Isolate vocals from music/noise using Meta's Demucs v4. Feed clean vocals into transcription.
batch_search	Search multiple audio files in parallel. Ideal for podcast libraries or interview collections.

Utilities

Tool	Description
text_to_speech	Convert text to natural speech using Kokoro TTS. 54 voices, 9 languages. Runs in background.
list_files	List media files in a directory with size info.
list_memories	Browse all stored transcriptions by title, duration, and date.
memory_stats	View memory statistics (file count, total duration).
clear_memory	Clear the transcription memory to free disk space.
tag	Add, remove, or list tags on transcriptions. Broad topic categories for organizing memories.
highlights	Export the best moments from a transcription. Auto mode picks top moments; focused mode finds moments matching a topic.
visual	Extract visual context from video at moments that matter. Query, auto, manual, and assist modes. Frames saved to Obsidian vault.
rebuild_graph	Rebuild Obsidian graph view data for all transcriptions. Migrates files, computes wikilinks, generates MOC hubs.
spaces	Download or live-record X/Twitter Spaces. Start, check status, or stop recordings.

Usage Examples

Take notes from a video

"Take notes from https://youtube.com/watch?v=xxx"

The agent calls take_notes which downloads, transcribes, and returns formatted notes. One tool call does everything.

Search a podcast for topics

"Search this podcast for every mention of AI regulation" — provide the file path or URL.

The agent uses search_audio for exact keyword matches, or deep_search for semantic matches (finds relevant discussion even without exact words).

Transcribe and identify speakers

"Transcribe this meeting recording and tell me who said what"

The agent calls transcribe_audio then identify_speakers to label each segment by speaker.

Search across all transcriptions

"Search everything I've ever transcribed for mentions of funding"

The agent uses search_memory to search across all stored transcriptions without needing a file path.

Export a clip

"Clip the part where they talk about pricing"

The agent uses search_audio or deep_search to find the moment, then clip_export to extract just that segment.

Separate vocals from noisy audio

"This recording has music in the background, clean it up and transcribe"

The agent calls separate_audio to isolate vocals, then transcribe_audio on the clean vocals track.

Generate speech from text

"Read these notes aloud"

The agent calls text_to_speech to generate an MP3 with natural speech. Supports multiple voices and languages.

Note Styles

When using take_notes, the style parameter controls formatting:

Style	Description
tldr	Shortest possible summary. One screen. Bold key terms.
notes	Clean sections with nested bullets (default).
highlight	Notes with callout blocks for key insights and blockquotes with timestamps.
eye-candy	Maximum visual formatting — callouts, tables, checklists, blockquotes.
quiz	Multiple-choice questions with answer key.

Model Sizes

tiny is the default and handles nearly everything. Only use larger models for heavy accents, poor audio quality, or maximum accuracy needs.

Model	Speed	Accuracy
tiny	Fastest	Excellent (default)
base	Fast	Excellent
small	Medium	Superior
medium	Slow	Outstanding
large	Slowest	Maximum

File Paths

Augent reads and writes to these locations on your machine:

Path	Purpose
~/Downloads/	Default directory for downloaded audio files
~/Desktop/	Default directory for notes, clips, and TTS output
~/.augent/memory/transcriptions.db	SQLite database for persistent transcription memory
~/.augent/memory/transcriptions/	Markdown files for each stored transcription
~/.augent/config.yaml	User configuration (optional)
~/.augent/auth.json	Twitter/X authentication cookies for Spaces recording (optional, user-created)

If Obsidian is installed, visual frames are saved to the Obsidian vault's External Files/visual/ directory. The vault path is auto-detected from Obsidian's config.

Network Access

Network access is used for two purposes only:

1. Downloading media from user-provided URLs via yt-dlp and aria2c
2. Downloading ML models on first use (Whisper, sentence-transformers, pyannote, Demucs, Kokoro) from Hugging Face

No telemetry. No background network activity. No data is uploaded.

ML Dependencies

The augent[all] install includes these local ML components:

Component	Purpose	Size
faster-whisper	Speech-to-text transcription	~75MB (tiny model)
sentence-transformers	Semantic search, auto-tagging, chapter detection	~90MB
pyannote-audio	Speaker diarization	~29MB
Demucs v4	Audio source separation (vocals from noise)	~80MB
Kokoro	Text-to-speech (54 voices, 9 languages)	~200MB

All models run locally. None require API keys or cloud services.

Requirements

• Python 3.10+
• FFmpeg (audio processing)
• yt-dlp + aria2c (for audio downloads)

Links

• GitHub
• Documentation