Transcribe Review

🤖 AI Drafted (Not reviewed)

Second-pass QA of a prior transcription against the image

Tool id	transcribe_review
Category	vision
Uses a language model	yes
Needs a generative model	no
Runs over many items	yes
Item handling	batch
Structured output	no
Human-verified	yes

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

Options

Option	Type	Default	What it does
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.
language	string	auto	Language hint for the reviewer (es / en / auto).
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.
skip_if_artifact_exists	boolean	yes	Reuse a matching prior review artifact.
temperature	number	0.7	Creativity.
thinking_mode	string	off	Chain-of-thought reasoning depth. One of: off, short, medium, long.
vision_mode	string	llm	Vision engine. ‘auto’ picks based on the resolved provider: apple → Apple Vision OCR; anything else → LLM vision path. ‘kraken’ → on-device neural line segmentation (baselines/polygons, no text). One of: auto, apple, llm, kraken. (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.

You are reviewing a prior transcription of a manuscript image.

Your job: compare the prior transcription against the image, identify
errors, and produce a CORRECTED transcription.

Common error patterns to look for:
- Abbreviations the prior pass missed or expanded incorrectly (Vmd, dho/dha, q̄, tpo, mrd, qta).
- Confused glyphs in cursive scripts: c↔e, rn↔m, u↔n, long-s mistaken for f, t↔l.
- Marginalia, stamps, signatures, or rubrics the prior pass skipped.
- Crossed-out text the prior pass either ignored or transcribed without marking.
- Inserted words above the line the prior pass dropped.
- Modernised spellings the prior pass introduced ('haver' silently turned into 'haber').
- Numbers and dates: digit confusions (1/7, 0/o, 5/6 in old hands).
- Names misread because the model fell back on a more familiar spelling.

Rules:
- Output ONLY the corrected transcription. No commentary, no diff, no
  notes about what you changed, no preamble.
- Preserve original orthography of the manuscript, including accents and
  diacritics, verbatim. Do NOT modernise.
- Keep [ilegible], [uncertain], [tachado: ...], [rúbrica], [sin texto]
  conventions.
- If the prior transcription is already correct, return it unchanged
  verbatim. Do not paraphrase or reformat.
- Do not invent text that is not visibly present.

The prior transcription appears in the Context section above. The image
follows.