Similarity

🤖 AI Drafted (Not reviewed)

Score image similarity


Tool id	`similarity`
Category	vision
Uses a language model	yes
Needs a generative model	yes
Runs over many items	no
Item handling	batch
Structured output	yes
Human-verified	not yet

What it reads

Port	Type	Required	What it is
Files (`files`)	`files`	yes	Image files
Context (`context`)	`any`	no	Previous text/transcription
Metadata (`metadata`)	`json`	no	Existing metadata
Documents (`documents`)	`json`	no	Document metadata

What it emits

Port	Type	Required	What it is
Text (`text`)	`text`	—	Raw text response
Value (`value`)	`any`	—	Parsed value
Texts (`texts`)	`array`	—	Per-item texts
Values (`values`)	`array`	—	Per-item values
Results (`results`)	`json`	—	Full results
Records (`records`)	`array`	—	Per-document text records [{doc_id, text}, …].
Artifacts (`artifacts`)	`json`	—	Artifact IDs
Clusters (`clusters`)	`json`	—	Typed same-document clusters.

Options

Option	Type	Default	What it does
`aspects`	array	[“content”, “composition”, “color”, “style”]	Aspects to score.
`choices`	array	—	Valid choices. (Not shown in the editor.)
`chunk_size_chars`	integer	0	Chunk large input text above this character budget (0=auto).
`force_ocr`	boolean	no	Force image processing instead of existing text.
`match_mode`	string	prefer	Match mode. One of: prefer, strict, inform.
`max_image_dimension`	integer	8192	Max image size.
`max_items`	integer	10	List max items.
`max_tokens`	integer	8192	Max response.
`max_words`	integer	50	Word limit.
`metadata_field`	string	—	Save to field.
`model_name`	string	—	Model name.
`output_format`	string	text	Response format. One of: text, boolean, choice, number, words, list, json.
`prompt`	string	—	Custom prompt.
`provider_name`	string	—	LLM provider. One of: openai, anthropic, google, ollama, lmstudio, groq, together, deepseek, mistral, openrouter, dashscope, xai, perplexity, fireworks, deepl.
`quality_gate`	boolean	yes	Stop the run if output is unreadable.
`reference_values`	object	—	Known values to match. (Not shown in the editor.)
`save_to_db`	boolean	yes	Save to library.
`save_to_file`	boolean	no	Export to file.
`scale`	string	percentage	Score scale. One of: percentage, 1-10, 1-5.
`temperature`	number	0.7	Creativity.
`thinking_mode`	string	off	Chain-of-thought reasoning depth. One of: off, short, medium, long.
`vision_mode`	string	auto	Vision engine. ‘auto’ picks based on the resolved provider: apple → Apple Vision OCR; anything else → LLM vision path. One of: auto, apple, llm. (Not shown in the editor.)

The prompt it sends

This is what the tool asks a model, with every option left at its default. Changing the options above changes this text.

Compare these images and score their similarity.

Score each aspect as a percentage (0-100, where 100 = identical)

Score these aspects: content, composition, color, style

Also provide an overall similarity score.

Also group images that are the same underlying document (duplicates, alternate scans,
alternate photos, near-identical variants). Consider both visual appearance and visible
text content. Every image must appear in exactly one same-document cluster.

Return as JSON:
{
    "overall_similarity": <score>,
    "aspect_scores": [
        {"aspect": "<aspect>", "score": <score>}
    ],
    "most_similar": "<which aspect is most similar>",
    "most_different": "<which aspect is most different>",
    "notes": "<brief explanation of key differences>",
    "same_document_clusters": [
        {
            "cluster_id": "cluster-1",
            "member_indexes": [0, 1],
            "similarity_score": 0.98
        }
    ]
}

In same_document_clusters, similarity_score is a fraction between 0 and 1
(NOT a percentage), regardless of the aspect score scale above.

Return ONLY valid JSON.