Skip to content
cccskills
FreePublished on GitHub

vision-audio

vision-audio: screenshot, describe, transcribe, diagram.

The English summary is being prepared. Showing the original description.

See how to install

Files included (1)

  • SKILL.md3.6 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Vision + Audio Pipeline — Multimodal Understanding

Beschreibung

scripts/vision-audio.py ist ein CLI-Tool für:

  • Screenshot aufnehmen und analysieren (--screenshot)
  • Bildbeschreibung via Pixelanalyse/Pillow (--describe)
  • Code-Diagramm-Screenshots lesen (Kontrast-basiert) (--diagram)
  • Audio-Transkription via OpenAI Whisper (--transcribe)
  • Mikrofon-Aufnahme + Transkription (--record)

Alle Ausgaben erfolgen als strukturiertes JSON.

Abhängigkeiten

pip install mss PyAudio openai-whisper Pillow

CLI Usage

usage: vision-audio.py [-h] [--screenshot] [--describe BILD]
                       [--diagram DIAGRAMM] [--transcribe AUDIO] [--record]
                       [--seconds SECONDS] [--json]

Vision + Audio Pipeline — Multimodal Understanding

options:
  -h, --help           show this help message and exit
  --screenshot         Screenshot aufnehmen + beschreiben
  --describe BILD      Bilddatei beschreiben (Pixel-Analyse)
  --diagram DIAGRAMM   Code-Diagramm-Screenshot analysieren
  --transcribe AUDIO   Audiodatei transkribieren
  --record             Mikrofon aufnehmen + transkribieren
  --seconds SECONDS    Aufnahmedauer in Sekunden (Default: 10)
  --json               Nur JSON ausgeben (keine Stderr-Info)

Beispiele

# Screenshot + beschreiben
python scripts/vision-audio.py --screenshot

# Bild beschreiben
python scripts/vision-audio.py --describe screenshot.png

# Code-Diagramm analysieren
python scripts/vision-audio.py --diagram code-flow.png

# Audio transkribieren
python scripts/vision-audio.py --transcribe meeting.wav

# Mikrofon aufnehmen + transkribieren
python scripts/vision-audio.py --record
python scripts/vision-audio.py --record --seconds 30

Architektur

vision-audio.py
├── Vision
│   ├── vision_screenshot()   → MSS / PowerShell-Fallback
│   ├── vision_describe()     → Pillow Pixelanalyse + Farbextraktion
│   └── vision_diagram()      → Kontrast-basierte Text-Bereichs-Detektion
├── Audio
│   ├── audio_transcribe()    → openai-whisper (Model: base)
│   └── audio_record()        → PyAudio → WAV → whisper
└── CLI (argparse)

Output Format (JSON)

{
  "action": "describe|screenshot|transcribe|diagram|record",
  "status": "ok|error",
  "timestamp": "2026-08-22T00:40:21.551751",
  "text": "Bild-Deskription / Transkription / Diagramm …",
  "file_path": "path/to/file.png",
  "metadata": {
    "width": 1920,
    "height": 1200,
    "format": "PNG",
    "mode": "RGB",
    "avg_brightness": 0.635,
    "contrast": 0.883,
    "total_pixels": 26666,
    "dominant_colors": [["rgb(240,240,245)", 820], ...],
    "thumbnail_base64": "iVBOR..."
  }
}

Tips & Pitfalls

  1. Whisper-Modell: Beim ersten --transcribe lädt Whisper das base-Modell herunter (~150 MB).
  2. Mikrofon: PyAudio benötigt ein funktionierendes Mikrofon. Fallback via PowerShell, falls PyAudio fehlt.
  3. Screenshot: mss (MSS) ist Standard; bei Fehler Fallback auf PowerShell PrintScreen.
  4. Code-Diagramm: Die Diagramm-Analyse ist heuristisch (Kontrast-Erkennung) – für echten OCR empfehle ich Tesseract.
  5. Bildbeschreibung: Sampling über Grid (40×40 Punkte) – gute Balance zwischen Performance und Genauigkeit.
  6. Ausgabeverzeichnis: Screenshots und Aufnahmen liegen unter ~/.openamer/vision-audio/.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Build integrated IS/BS/CF financial workbooks in Excel.

The English summary is being prepared. Showing the original description.

openamer/openamer8Updated Oct 21, 2026

Run and grow the OpenAmer Agent-to-Agent swarm: identity, trust, node-to-node ask, signed skill/insight sharing, and the autonomous self-learning loop.

The English summary is being prepared. Showing the original description.

openamer/openamer8Updated Oct 21, 2026

Use for A/B experiments on OpenAmer configs and skills.

The English summary is being prepared. Showing the original description.

openamer/openamer8Updated Oct 21, 2026

Roleplay a hostile user to find and triage UX pain points.

The English summary is being prepared. Showing the original description.

openamer/openamer8Updated Oct 21, 2026

Agent mesh: master/worker nodes, HTTP delegation, heartbeat.

The English summary is being prepared. Showing the original description.

openamer/openamer8Updated Oct 21, 2026

Use for agentic commerce payments tasks. Candidate synthesized from trend radar; awaiting trial promotion.

The English summary is being prepared. Showing the original description.

openamer/openamer8Updated Oct 21, 2026

See all skills by openamer

Report a problem with this skill