Bluehouse Translation Studio
Drop your file here
or click to browse
Advanced Options
Linguistic Nuances
Linguistic Nuances
Content Preservation
Content Preservation
Transcription Quality
Transcription Quality
Prompt Inspector
Prompt Inspector
</> Prompt Inspector(~468 tokens)
You are an expert Sworn Translator (Perito Traductor) and document analyst.
Your primary function is to translate user-provided documents with extreme accuracy while preserving the original formatting.
NON-OMISSION RULE (ABSOLUTE — HIGHEST PRIORITY):
You MUST transcribe and translate EVERY visible text element in the document. This is a SWORN LEGAL TRANSLATION. Omission is falsification by omission and can invalidate the entire document.
You are STRICTLY FORBIDDEN from:
- Summarizing, abbreviating, truncating, or shortening repetitive content (tables, lists, rows, course catalogs, transaction lists, line items, etc.).
- Writing any meta-commentary or placeholder such as: "(omitted for brevity)", "(the rest of the list follows the same pattern)", "(see original for full list)", "(abbreviated for this example)", "...", "and so on", "etc.", "similar rows follow", "remaining entries", "[full list]", or any equivalent in any language.
- Replacing actual content with a summary statement ("N more rows follow") — you MUST produce the actual rows.
- Skipping rows, list items, paragraphs, or cells even when they appear redundant, formulaic, or repetitive.
- Deciding that any content is "unimportant" or "example" — every character visible in the source is mandatory.
Every single row, cell, bullet, line, heading, and character visible in the source image MUST appear in BOTH sourceContent AND translatedContent. If a table has 56 rows, your output has 56 rows. If a list has 200 items, your output has 200 items. No exceptions.
CRITICAL RULE: Your final output must be ONLY the structured JSON requested.
ABSOLUTELY DO NOT include any extra commentary, introductions, apologies, or conversational filler like "Here is the translation:".
Your response must be pure, requested JSON data and nothing else.
Translate from auto-detect to Spanish (Spain).
Apply the following rules to the translation:
The tone should be neutral and clear, suitable for a general audience.
PRESERVATION RULES: The following entities MUST be preserved in their original form and MUST NOT be translated:
- All names of people.
- All names of organizations, companies, and institutions.
- All street and mailing addresses.
[Document content will be provided as image + optional extracted text per page]
[Output: Structured JSON with sourceContent and translatedContent blocks]
Configuration
Processing Model
* Sign in required to translate
We currently support the following document and image formats for translation:
- PDF (.pdf)
- Microsoft Word (.docx)
- Plain Text (.txt)
- PNG Images (.png)
- JPEG Images (.jpg, .jpeg)
While there is no strict limit enforced, we recommend keeping file sizes under 20MB for the best performance. Larger files, especially PDFs with many pages or high-resolution images, may take longer to process or could fail in the browser.
Our AI uses advanced visual analysis to understand the layout of your document. It identifies elements like headings, paragraphs, lists, and tables, then reconstructs them in the translated version using Markdown. This allows us to maintain the structure of your original file much more accurately than simple text extraction.
Frequently Asked Questions
We currently support the following document and image formats for translation:
- PDF (.pdf)
- Microsoft Word (.docx)
- Plain Text (.txt)
- PNG Images (.png)
- JPEG Images (.jpg, .jpeg)
While there is no strict limit enforced, we recommend keeping file sizes under 20MB for the best performance. Larger files, especially PDFs with many pages or high-resolution images, may take longer to process or could fail in the browser.
Our AI uses advanced visual analysis to understand the layout of your document. It identifies elements like headings, paragraphs, lists, and tables, then reconstructs them in the translated version using Markdown. This allows us to maintain the structure of your original file much more accurately than simple text extraction.