Skip to content

saravanan-ui/humanizer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Humanizer v3.4.0

Remove AI writing patterns from B.Tech project reports — and any other writing.

Built for CSE/IT students writing their final year project reports. Works for essays, articles, and documentation too.


The problem

Most B.Tech project reports open with:

"In today's rapidly evolving technological landscape, this project aims to develop a proposed system that enhances user experience..."

Examiners have read that sentence ten thousand times. It signals immediately that the student didn't write this — and they stop reading carefully.

The Humanizer fixes that. It strips 24 documented AI writing patterns and replaces them with writing that sounds like you actually built something.


What it fixes

Report-specific patterns:

  • Generic openers ("In today's rapidly evolving technological landscape...")
  • Passive voice overload ("it was observed that", "testing was conducted")
  • Vague performance claims ("the system performed well") — replaced with real numbers
  • "The proposed system" — replaced with what you actually built
  • Literature review that just lists papers without connecting them to your problem
  • Abstract that describes the report structure instead of the findings
  • Conclusion that repeats the introduction

General AI writing patterns:

  • Overused AI vocabulary (crucial, pivotal, leverages, showcases, underscores...)
  • Em dash overuse
  • Inline-header bullet lists
  • Forced rhetorical structure (rule of three, negative parallelism)
  • Promotional language (nestled, breathtaking, groundbreaking...)
  • Vague attributions (experts argue, industry observers...)
  • Sycophantic tone and chatbot artifacts
  • Excessive hedging and filler phrases
  • Synonym cycling and false ranges
  • And 9 more — see the full file

Files in this repo

File What it is What to do with it
humanizer-v3.4.0.md Full humanizer instructions Upload to your Claude project
student-setup-guide.html Step-by-step setup guide with UI mockups Open in browser
README.md This file Read it
LICENSE MIT License Nothing, just legal

Setup — 5 minutes, free

Step 1 — Create a Claude.ai account

Go to claude.ai and sign up. The free tier is enough for an entire project report.

Step 2 — Create a new Project

In the left sidebar, click ProjectsNew project. Name it something like B.Tech Report.

Step 3 — Upload the humanizer file

Inside your project, find the Files section. Click + and upload humanizer-v3.4.0.md.

Step 4 — Add the standing instructions

Click + next to Instructions. Paste this exactly:

You are a humanizer editor. Apply the humanizer checklist to all writing.
- For B.Tech reports and long-form writing: show draft → audit → final pass
- For short replies: apply silently, just write clean
- For B.Tech reports: every factual claim needs a real number or source
- At session start: ask the user to upload their project synopsis before writing anything
- Full instructions are in the uploaded file humanizer-v3.4.0.md — read it at session start

Step 5 — Test it

Start a new conversation inside the project. Claude will ask for your project synopsis. Upload it. Then paste any report section and say which chapter it is.

You get: draft → audit → final pass — with a checklist of anything that still sounds AI-generated.


How to use it

For each report section:

  1. Start a new conversation inside your Claude project
  2. Upload your project synopsis when Claude asks
  3. Paste the section you want to humanize
  4. Tell Claude which section it is: "This is my Introduction. Humanize it."
  5. Get draft → audit → final output
  6. Copy the final version into your report

For professors

Download both files and share with your batch:

  • humanizer-v3.4.0.md — students upload to their Claude project
  • student-setup-guide.html — students open in browser for the full visual setup guide

The instructions auto-configure to each student's project from their uploaded synopsis. No editing needed between students or projects.


How it works — the session start protocol

At the start of every conversation, Claude asks for your project synopsis (2-3 pages). It reads the document silently, extracts your project title, tech stack, methodology, and key metrics, then confirms before starting.

This means:

  • You never type out project details manually
  • Every humanized section uses language specific to your actual project
  • The instructions work for any project without editing

Version history

Version Changes
v3.4.1 B.Tech report quality checklist, preliminary pages section, formatting rules, acknowledgements protocol, clarity fixes
v3.4.0 Synopsis upload session start, fully generic (no hardcoded project names)
v3.3.0 Three-question session start protocol
v3.2.0 Reviewer persona, "did you actually build this" checklist, red flags quick reference
v3.1.0 B.Tech report mode, section tone calibration, report section template
v3.0.0 Soul/personality layer, second voice example, merged overlapping patterns
v2.3.0 Initial release

License

MIT — free to use, share, and modify. See LICENSE for details.

Patterns adapted from Wikipedia:Signs of AI writing (CC BY-SA 4.0). Original instructions, B.Tech additions, and session protocol are original work.

About

24-pattern checklist for B.Tech reports. Runs on Claude.ai. Catches passive voice, vague claims, generic conclusions, and 21 more tells.

Topics

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages