Most tools for checking plagiarism or AI-generated content share a few drawbacks:
- they require uploading your text to a website
- they are often paywalled or rate-limited
- they lack transparency in how results are generated
I ran into this problem a few times and wanted a simpler alternative I could run locally.
So I built dokimos — a lightweight Python CLI for checking:
- text similarity (potential plagiarism)
- AI-likeness signals
What It Looks Like
$ dokimos check essay.txt
Plagiarism risk: Low (12%)
AI-likeness: Medium (0.63)
The goal is not to replace large commercial tools, but to provide something that is:
- fast
- minimal
- scriptable
- privacy-friendly (no text leaves your machine)
Why I Built This
I wanted a tool that could:
- quickly validate content before submitting or sharing
- integrate into scripts or developer workflows
- work entirely offline
Many existing solutions feel heavy or opaque. This is intentionally simple and local-first.
How It Works (High-Level)
At the moment, dokimos focuses on:
- similarity checks to flag potentially copied content
- heuristic-based indicators to estimate AI-generated patterns
It is still an early-stage project, and the results should be treated as signals rather than definitive conclusions.
Target Use Cases
- developers who prefer CLI-based tools
- students and writers who want a quick local check
- teams looking for simple automation hooks
Roadmap
Planned improvements include:
- better detection accuracy
- clearer explanations for scores
- structured output (e.g., JSON) for integration
- deeper similarity insights
Repository
https://github.com/Thavarshan/dokimos
Feedback
This is an early project, so feedback is very valuable.
If you have suggestions around accuracy, usability, or features, I would appreciate hearing your perspective.
Top comments (1)
Running plagiarism and AI-detection locally is a great privacy-preserving approach. No need to upload sensitive academic work to third-party servers. Useful for educators and students alike!