Image Recognition

Use this page to configure Groq Image Recognition services for camera snapshots, Home Assistant media images, allowlisted local files, image URLs, and OCR-style prompts.

Contents

• Setup
• Analyze image action
• Extract text action
• Image inputs
• Built-in fallback models
• Related pages

Setup

1. Open the Groq integration page.
2. Choose Add Image Recognition.
3. Enter a friendly service name.
4. Select a vision-capable model.
5. Optionally set a system prompt to guide answers.
6. Enable free-tier protection if you want this service paused locally after Groq reports a limit reset time.

View	Screenshot
Image Recognition setup
Analyze Image action

Analyze image action

Use groq.analyze_image when you want the model to answer a question about an image. The action requires the Image Recognition service_id.

action: groq.analyze_image
data:
  service_id: Front Camera Vision
  prompt: Describe anything unusual in this camera image.
  camera_entity_id: camera.front_yard
  model: meta-llama/llama-4-scout-17b-16e-instruct
response_variable: groq_image

You can also pass an image URL:

action: groq.analyze_image
data:
  service_id: Front Camera Vision
  prompt: Describe anything unusual in this camera image.
  image_url: https://example.com/snapshot.jpg
response_variable: groq_image

Extract text action

Use groq.extract_text_from_image when you want OCR-style output from the configured Image Recognition service. The action requires the Image Recognition service_id.

action: groq.extract_text_from_image
data:
  service_id: Entry OCR
  camera_entity_id: camera.entry_table
  prompt: Extract all visible text. Return only the text.
response_variable: groq_ocr

Image inputs

The image actions support these input sources:

Input	Use it for	Notes
camera_entity_id	Current Home Assistant camera snapshot	Best for automations and security/event workflows.
image_file	Home Assistant media image	Use the media selector from Developer Tools -> Actions.
image_path	Advanced local file path	The path must be allowed in Home Assistant allowlist_external_dirs.
image_url	Public or data URL	Useful for webhook or externally hosted snapshots.

Provide one image source per call.

Built-in fallback models

The setup flow prefers live Groq model discovery. If discovery is unavailable, the integration has built-in fallback model IDs:

• meta-llama/llama-4-scout-17b-16e-instruct
• qwen/qwen3.6-27b
• meta-llama/llama-4-maverick-17b-128e-instruct

The model registry also detects Qwen vision models with qwen/qwen3.6- IDs and vision-capability metadata returned by Groq model discovery.

Related pages

• Account and Service Setup
• Text Generation
• Troubleshooting