In an era where algorithms can identify an author based purely on their sentence length, punctuation habits, and vocabulary choices, maintaining privacy goes far beyond hiding your IP address. Enter **IAnon**, a powerful new Windows application designed to disrupt stylometric analysis and protect author identity by masking distinctive linguistic patterns while keeping the core meaning intact.
For readers of *veydunet.com* following the intersection of open-source software and digital privacy, IAnon represents a fascinating shift: bringing AI-powered stylometric obfuscation directly to the local machine.
The Threat of Stylometry: Your Writing Has a Fingerprint
Every writer leaves a unique linguistic footprint. Stylometry—the statistical analysis of authorship—examines recurring surface characteristics such as average sentence length, paragraph structures, vocabulary diversity, and connector frequency. Even when names, metadata, and IP addresses are stripped away, these micro-habits can deanonymize an author.
While traditional anonymization tools focus on network traffic or metadata, IAnon targets the text itself. Developed as part of the **Ostium** toolset (alongside the Ostium OSINT Browser), IAnon tackles the problem head-on by rewriting text locally to neutralize these distinctive patterns.
(Note: The developers emphasize an important caveat: while IAnon makes stylometric analysis significantly more difficult, it does not guarantee absolute, 100% anonymity.)
Under the Hood: Local AI Meets C# Style Analysis
What sets IAnon apart from generic cloud-based rewriting tools is its strict commitment to local execution and architectural transparency.
1. Local Style Profiling
Before sending anything to an LLM, IAnon performs a local **Style Analysis** phase written entirely in C#. Without needing external requests, it evaluates lightweight linguistic metrics:
- Average sentence length and its variance
- Vocabulary diversity and repeated wording
- Punctuation patterns and connector frequencies
- A baseline formality estimate
This generates a precise transformation profile. Instead of asking an AI to perform a blind, generic rewrite, IAnon feeds these explicit structural targets directly into the transformation prompt.
2. Powered by Ollama
IAnon runs locally via **Ollama**, communicating directly with its HTTP API (`/api/tags` and `/api/chat`) using a native `HttpClient` with no bloated SDK dependencies. Users can choose their preferred model, select target languages, styles, and syntax variations, and watch the transformation happen offline.
Privacy-First Architecture
True to its mission, IAnon’s design philosophy places user sovereignty at the forefront:
- Loopback Default: The Ollama endpoint defaults securely to loopback (
127.0.0.1), with remote endpoints disabled out of the box. - Zero Disk Footprint: Source and output texts are never written to disk by the application.
- No Telemetry: The app contains zero analytics, cloud tracking, or telemetry mechanisms.
- Full Customization: Prompts, style descriptions, and syntax variations are entirely editable via
config.jsonor the application’s settings menu. No API keys are required.
Getting Started with IAnon
Running IAnon requires a Windows 10/11 environment, the .NET 10 Desktop Runtime/SDK, and a running instance of Ollama with at least one model installed.
You can build it from source using PowerShell:
dotnet restore
dotnet build -c ReleaseOr publish it as a self-contained single file:
dotnet publish -c Release -r win-x64 --self-contained true -p:PublishSingleFile=trueOnce running, the workflow is streamlined: start Ollama, launch IAnon, select your desired language and style variations, input your text, click **ANALYZE STYLE**, and hit **ANONYMIZE**.
Released under the permissive **MIT License**, IAnon is available now on GitHub for privacy advocates, researchers, and writers looking to reclaim control over their digital prose.