# AI use cases for Humanizer Local Model (sgaofen/humanizer-local-model) Repo: [sgaofen/humanizer-local-model](https://github.com/sgaofen/humanizer-local-model) | Use cases: 4 | Page: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/ A 12B local model that rewrites AI-written drafts so they read like a person wrote them, keeping numbers, names, dates and quotes, and running on your own computer. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Rewrite an AI-drafted email offline before sending it: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/rewrite-ai-email-before-sending/ - Rewrite a long Markdown or Word document in one command: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/rewrite-long-document-with-hz/ - Rewrite a report while keeping chosen quotes word for word: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/rewrite-report-keeping-exact-quotes/ - Let a coding agent batch-rewrite drafts through a local server: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/agent-batch-rewrite-via-local-server/ ## How to rewrite an AI-drafted email offline before sending it with Humanizer Local Model Repo: [sgaofen/humanizer-local-model](https://github.com/sgaofen/humanizer-local-model) | Difficulty: Beginner | Time: 45 minutes | Basis: From the README Paste an AI-written email into a local app and get a version that reads like a person wrote it. **Who it is for:** An office worker or student who has an AI-drafted email that sounds generic and does not want to send a draft that reads as machine-written. **The problem:** AI drafts often sound stiff or templated, and the sender has to rewrite them by hand without losing the facts. **The outcome:** A rewritten email with new wording highlighted and replaced wording struck through, produced on your own computer after a one-time model download. ### Steps 1. Download the app for your computer from the Quick start section: the Mac (Apple silicon) .dmg or the Windows x64 setup .exe. 2. Open the app. Expect a first-launch warning, since the app is not code-signed yet; the one-time fix is in docs/INSTALL.md. 3. Let the app download the suggested model size once from Hugging Face. After that it works offline. 4. Paste the email draft on the left. Read the rewrite on the right. 5. Check every number, date and name against the original before you copy the rewrite into your mail client. ### Prompt to paste into your agent ```text Using the sgaofen/humanizer-local-model repo, install the app for this machine and rewrite the email draft at with the app. Expected output: the rewritten email text. Verify that every number, date and name in the draft also appears unchanged in the rewrite. ``` ### Caveat The README says the model still changes details in some rewrites, so check numbers, dates and names before sending. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/rewrite-ai-email-before-sending/ ## How to rewrite a long Markdown or Word document in one command with Humanizer Local Model Repo: [sgaofen/humanizer-local-model](https://github.com/sgaofen/humanizer-local-model) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Use the hz command line tool to rewrite a whole article or .docx while keeping its structure. **Who it is for:** A researcher, writer or team member who has a long AI-drafted paper, article or report file and wants the prose rewritten without redoing headings, tables or links by hand. **The problem:** Rewriting a long file piece by piece is slow, and it is easy to lose a number or a structural element along the way. **The outcome:** A rewritten output file with headings, code blocks, tables and links kept as they were, plus a per-piece report of any number that went missing or was added. ### Steps 1. Install the app from Option 1 or set up llama-server from Option 2 so a local model is available. 2. Run `pipx install git+https://github.com/sgaofen/humanizer-local-model` (or `pip install git+https://github.com/sgaofen/humanizer-local-model`). 3. For a .docx file, run `pipx inject humanize-model python-docx`. 4. Run `hz paper.md -o paper.out.md` (replace the file names with yours). 5. Read the warnings on stderr for any missing or new number, then read the whole output, since names and meaning are not checked. ### Prompt to paste into your agent ```text Using the sgaofen/humanizer-local-model repo, install the hz tool with pipx and run `hz -o ` on the target document. Expected output: a rewritten file with structure preserved. Verify that the stderr report lists no lost or added numbers, or that each listed piece has been checked by hand. ``` ### Caveat The README warns that in a .docx, bold or italic inside a paragraph is lost, and that names and the meaning of sentences are not checked. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/rewrite-long-document-with-hz/ ## How to rewrite a report while keeping chosen quotes word for word with Humanizer Local Model Repo: [sgaofen/humanizer-local-model](https://github.com/sgaofen/humanizer-local-model) | Difficulty: Beginner | Time: 45 minutes | Basis: From the README Import a Word or PDF report and mark the passages that must stay exactly as written. **Who it is for:** A policy analyst or project lead who has an AI-drafted report with exact quotes or figures that must not change. **The problem:** A general rewrite can alter a quote or a figure, and the author needs certain passages to stay verbatim. **The outcome:** A rewritten report where the selected passages appear unchanged and are highlighted in both the draft and the rewrite. ### Steps 1. Install the app for your computer from the Quick start section and open it. 2. Import the text of your Word (.docx) or PDF file using the app. 3. Select the passage that must stay exact, then click Create fact. 4. Click rewrite and wait for the result; check that the kept passages are highlighted in the draft and the rewrite. 5. Edit the draft if needed and run again, since the draft stays editable after a rewrite. ### Prompt to paste into your agent ```text Using the sgaofen/humanizer-local-model repo app, import and mark the passages listed in with Create fact before rewriting. Expected output: a rewritten report in which each marked passage appears word for word. Verify that each marked passage is highlighted in the rewrite and matches the original exactly. ``` ### Caveat The README says the model still changes details in some rewrites, so anything not marked with Create fact still needs checking. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/rewrite-report-keeping-exact-quotes/ ## How to let a coding agent batch-rewrite drafts through a local server with Humanizer Local Model Repo: [sgaofen/humanizer-local-model](https://github.com/sgaofen/humanizer-local-model) | Difficulty: Advanced | Time: 1 hour | Basis: From the README Run the 12B model as a local llama-server and have an agent send each draft to it. **Who it is for:** A developer or team building an automated pipeline that rewrites many AI-drafted texts and needs a local, scriptable endpoint. **The problem:** Rewriting many drafts by hand does not scale, and a cloud service may not suit private text. **The outcome:** A local llama-server answering rewrite requests, with an agent looping over a folder of drafts and saving each rewrite. ### Steps 1. Start the model with `llama-server --hf-repo jialinyyzz/humanizer --hf-file humanizer-12b-Q8_0.gguf -c 8192 -np 1 -ngl 99 --port 8080`. On a 16 GB machine, use humanizer-12b-Q6_K.gguf instead. 2. Download the prompt format with `curl -sLO https://huggingface.co/jialinyyzz/humanizer/resolve/main/prompt_format.json`. 3. Send each draft to `/completion` using the jq and curl pipeline in Option 2 of the README, not `/v1/chat/completions`. 4. Save each returned text to its own output file and keep the draft next to it for review. 5. Read docs/USAGE.md for the folder script and the troubleshooting table before you scale up. ### Prompt to paste into your agent ```text Using the sgaofen/humanizer-local-model repo, start llama-server with the Q8_0 GGUF on port 8080 and rewrite every .txt file in through the /completion endpoint with the prompt format from prompt_format.json. Expected output: one rewritten .txt per input file in . Verify that a single test draft returns non-empty text from the .content field before running the whole folder. ``` ### Caveat The README says this is a plain text-completion model, not a chat model, so it must be called through /completion with the stated prompt format. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/sgaofen-humanizer-local-model/agent-batch-rewrite-via-local-server/