Text Style Editor – User Documentation#

1. Purpose#

The Text Style Editor is a web-based application for the stylistic transformation of texts. It allows any source text to be converted into a predefined or individually configured language style.

The basic principle follows a two-step process:

1. Neutralisation (optional): Stylistic peculiarities are removed from the input text to create a uniform starting point.
2. Stylisation: The neutralised or original text is transformed into the target style according to the selected control settings.

The application is aimed at people who need to prepare texts for different communication contexts – such as academic papers, official communications, marketing purposes or creative projects.

2. Range of functions#

Core functions#

• Text transformation: Conversion of texts into different language styles using AI-supported language processing
• 34 style controls: Fine-grained control of stylistic features in seven thematic groups
• 23 presets: Predefined style combinations for common use cases
• Neutralisation: Removal of stylistic idiosyncrasies before the actual transformation
• History: Storage and comparison of the last 20 transformations
• Export settings: Saving and loading control configurations as shareable code

The seven control groups#

Control types#

• Polar: Opposite pairs with a scale from -10 to +10 (e.g. formal to informal)
• Intensity: Strength of a characteristic from 0 to 10 (e.g. irony, objectivity)
• Levels: Discrete options (e.g. language level A1, A2, B1, B2, C1, C2)

3. Operation#

Step-by-step instructions#

1. Enter text: Paste your source text into the input field. The token counter shows you the current text length.

2. Select preset (optional): Select a predefined style profile from the drop-down menu. The corresponding controls will be configured automatically.

3. Adjust controls: Open the desired control groups by clicking on them. Activate individual controls using the checkboxes and set the desired intensity.

4. Configure neutralisation (optional): Switch to the ‘Neutralisation’ tab to specify which stylistic aspects should be removed before the transformation.

5. Start transformation: Click on ‘Transform’. The progress is displayed.

6. Check result: The result appears in the output field. If neutralisation is active, you can view the intermediate text in the drop-down area.

Important controls#

Important notes#

• Token limits: The input text must contain at least 20 and no more than 10,000 tokens.
• Only activated controls have an effect: A control only influences the transformation if its checkbox is activated.
• Intensity levels: Values from 1–2 produce slight variations, values from 9–10 produce extreme variations of the style feature.
• Neutralisation is optional: You can deactivate this if you want to retain the original style.

4. Application example#

Scenario: Revision of a text#

Initial situation: An employee has written a draft introduction for a document. The content of the text is correct, but the style is too colloquial and too personal.

Procedure:

1. The person inserts the draft into the input field.

2. They select or create a preset, e.g.:

• Technical language (high)

• Hedging (clear)

• Structured (high)

• Personal (strongly negative, i.e. impersonal)

• Active (negative, i.e. passive constructions preferred)

• Language level C2

3. In the ‘Neutralisation’ tab, they leave the default settings (tonality, register and redundancy active).
4. After clicking on ‘Transform’, they receive a revised version that:

• Replaces colloquial expressions with technical terminology

• Converts first-person formulations into impersonal constructions

  • Inserts mitigating formulations (‘could’, ‘seems’)

• Removes filler words and redundancies

5. The person compares the original text with the transformed text in the history and specifically incorporates suitable formulations into their work.

5. Recommendations for efficient use#

General best practices#

• Start with a preset that is close to your target style and then adjust individual controls.
• Only activate the controls you actually need. Too many controls active at the same time can lead to inconsistent results.
• Use neutralisation if your source text is highly stylised (e.g. emotional or colloquial).
• Save frequently used configurations as code using the copy function.

Tips for optimal results#

• Shorter texts transform better: Divide long texts into meaningful sections.
• Use extreme values sparingly: intensities of 9–10 lead to highly exaggerated results that may require post-processing.
• Understand polar controls: for polar controls (e.g. formal/informal), a value of 0 means neutrality, while negative values reinforce the opposite.
• Use history: Compare different transformations of the same text to find the optimal settings.
• Use neutralisation selectively: Deactivate neutralisation if certain stylistic elements of the original are to be retained.

For scientific texts#

• Combine hedging with high precision for appropriate academic caution.
• Use the ‘Nominal’ setting for typical scientific nominal style.
• Select language level C2 for complex technical texts.

6. System limitations#

What the text style editor does not do#

• No content review: The system only transforms the style, not the accuracy of the content. Factual errors, logical inconsistencies or factual inaccuracies remain unchanged.
• No plagiarism check: The editor does not check whether the transformed text reproduces copyrighted material.
• No formatting: The system processes plain text. Complex document structures (tables, illustrations, footnotes) are not transferred.
• No translation: The text is output in the same language in which it was entered.
• No real-time correction: Spelling and grammar errors in the input text may be carried over into the output.

Technical limitations#

• Maximum text length: 10,000 tokens (equivalent to approximately 6,000–8,000 words)
• Minimum text length: 20 tokens
• Processing time: A few seconds to one minute, depending on text length and server load
• Session-based storage: The history is lost when the browser is closed

Quality notes#

• Conflicting slider settings (e.g. ‘Factual’ and ‘Emotional’ both set to high) can affect the output quality.
• Very short texts (less than 50 words) offer little scope for stylistic transformation.
• Technical terminology is not always correctly recognised or retained.

7. Summary#

The Text Style Editor is a tool for AI-supported stylistic text revision. It is particularly suitable for use cases in which texts need to be adapted to different communication contexts – from scientific terminology to official communications and creative styles.

The strength of the system lies in its fine-grained controllability through 34 individual controls and 23 predefined presets. The optional neutralisation level enables more consistent results for stylistically heterogeneous source texts.

The Text Style Editor does not replace human judgement. It is designed as a support tool whose results should always be critically reviewed and, if necessary, reworked. You remain responsible for the correctness of the content and stylistic appropriateness of the final product.