# Built my own “AI humanizer” tool after facing issues - need advice on scaling & launching 🚀

**URL:** <https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851>\
**Category:** Discussion & Solutions\
**Tags:** solved\
**Created:** [April 6, 2026, 6:54pm UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851 "2026-04-06T18:54:45Z")\
**Posts on this page:** 9\
**Page:** 1

<div class="post-metadata">

**Author:** ![reaper](https://onehack.st/letter_avatar_proxy/v4/letter/r/f04885/32.png) [@reaper](https://onehack.st/u/reaper)\
**Post date:** [April 6, 2026, 6:54pm UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/1 "2026-04-06T18:54:45Z")

</div>

Hey everyone,

I wanted to share a quick experience and get some guidance.

While working on my university assignments with the help of AI, things got really hectic. A friend suggested I try a tool called “ **Stealth Writer** ,” so he got subscription for me and i used it and got my work completed.

And here comes the second submission **Free version** It worked fine initially, but right when I had a submission deadline, their server went down and the tool was only available for paid users at that moment.

That situation pushed me to try building my own version of a “stealth writer” / AI humanizer.

So far, I’ve managed to build something that actually works. The text it generates is much more human-like, and when I test it with AI detection tools (including Stealth Writer’s detector), it usually shows low AI percentages often around 10%, and most results stay under 15–18%.

Now I’m at a point where I’d really appreciate some guidance:

- How can I improve this further to make the output even more natural and reliable?

- What are the best ways to scale something like this technically?

- How should I deploy it (web app, API, etc.)?

- And most importantly, what’s the best way to launch and get this in front of other users?

I’d love to hear your thoughts, suggestions, or even similar experiences.

Thanks in advance 🙌

---

<div class="post-metadata">

**Author:** ![Cesar\_Santos](https://onehack.st/user_avatar/onehack.st/cesar_santos/32/143229_2.png) [@Cesar\_Santos](https://onehack.st/u/Cesar_Santos)\
**Post date:** [April 6, 2026, 9:31pm UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/2 "2026-04-06T21:31:50Z")

</div>

Hi, congrats for your hard work.

Some questions here:

What Language it was built?

if node.js, Python,PHP or any other “web friendly” language it can be deployed easily via docker and a common web hosting

Do you want deploy it as a open source app?

If yes, the best approach is create a Github account with apache 2.0 license and deploy the code, then if was built in some web language like node js or python, deploy with github codespace to start, and scale up with a public cloud like AWS or GCP if you need heavy power or some hosting platform like Hostinger if you need a moderate use, even a free tool like vercel may be use (at free tier with some limitation)

about improove the output, I dont have much experience with it, I’ve customized some LLM at my work, but nothing so specialized like your.

---

<div class="post-metadata">

**Author:** ![reaper](https://onehack.st/letter_avatar_proxy/v4/letter/r/f04885/32.png) [@reaper](https://onehack.st/u/reaper)\
**Post date:** [April 7, 2026, 2:39am UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/3 "2026-04-07T02:39:15Z")

</div>

Hey, thanks a lot I really appreciate it! 🙌

Right now, I built it using **Python + Node.js** , mainly focusing on getting the core rewriting logic working first. I haven’t fully optimized the deployment side yet.

About open source I’m still deciding. Part of me wants to open it up and get community feedback, but I’m also thinking about turning it into a small product first and seeing how it performs. Maybe I’ll start closed and later release a limited or lite version as open source.

Also good point on GitHub Codespaces I hadn’t really considered that for initial setup.

For improving output, yeah it’s been a lot of trial and error with prompts and rewriting strategies so far. I’m now trying to make it more consistent and less “pattern-like,” since that’s where AI text still feels obvious.

Thanks again for taking the time to share all that it really helped! 👍

---

<div class="post-metadata">

**Author:** ![Cesar\_Santos](https://onehack.st/user_avatar/onehack.st/cesar_santos/32/143229_2.png) [@Cesar\_Santos](https://onehack.st/u/Cesar_Santos)\
**Post date:** [April 8, 2026, 12:56pm UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/4 "2026-04-08T12:56:09Z")

</div>

With Python and node the default pipeline is create a [CI/CD with Docker](https://docs.docker.com/compose/gettingstarted/) and then deploy it, even if you intend create a Saas subcription based software, you can create a container and use a hosting solution, your project is web friendly and can be deployed without complications, a extra tip is if possible create two separate projects one to your frontend and another to backend, it will help when you need modify a feature and allow isolations at docker deploy pipeline.

---

<div class="post-metadata">

**Author:** ![reaper](https://onehack.st/letter_avatar_proxy/v4/letter/r/f04885/32.png) [@reaper](https://onehack.st/u/reaper)\
**Post date:** [April 8, 2026, 1:51pm UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/5 "2026-04-08T13:51:42Z")

</div>

okay will try doing that, also the beta release of it. Thanks for being helpful…

---

<div class="post-metadata">

**Author:** ![Tempo\_Khan](https://onehack.st/user_avatar/onehack.st/tempo_khan/32/76599_2.png) [@Tempo\_Khan](https://onehack.st/u/Tempo_Khan)\
**Post date:** [May 8, 2026, 2:13am UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/6 "2026-05-08T02:13:03Z")

</div>

> [@reaper](#):
>
> Hey everyone,
> 
> I wanted to share a quick experience and get some guidance.
> 
> While working on my university assignments with the help of AI, things got really hectic. A friend suggested I try a tool called “ **Stealth Writer** ,” so he got subscription for me and i used it and got my work completed.
> 
> And here comes the second submission **Free version** It worked fine initially, but right when I had a submission deadline, their server went down and the tool was only available for paid users at that moment.
> 
> That situation pushed me to try building my own version of a “stealth writer” / AI humanizer.
> 
> So far, I’ve managed to build something that actually works. The text it generates is much more human-like, and when I test it with AI detection tools (including Stealth Writer’s detector), it usually shows low AI percentages often around 10%, and most results stay under 15–18%.
> 
> Now I’m at a point where I’d really appreciate some guidance:
> 
> - How can I improve this further to make the output even more natural and reliable?
> - What are the best ways to scale something like this technically?
> - How should I deploy it (web app, API, etc.)?
> - And most importantly, what’s the best way to launch and get this in front of other users?
> 
> I’d love to hear your thoughts, suggestions, or even similar experiences.
> 
> Thanks in advance

**Hey man, solid hustle turning that frustration into your own tool.**  
Congrats on already getting it down to 10-18% on detectors, that’s honestly better than a lot of the paid ones out there.

### On making it even more natural/reliable:

- **Vary sentence rhythm and structure heavily.** Humans ramble, use fragments, contractions, and sometimes repeat ideas for emphasis. Train or prompt it to occasionally add personal asides, transitional fluff (“honestly though…”, “the thing is…”, etc.).
- **Inject micro-imperfections.** Light typos (that you can toggle), colloquialisms, regional slang if it fits the audience, or slight redundancy. But keep it controlled—nothing that screams “I’m trying too hard.”
- **Multi-pass approach.** Generate → humanize → rewrite specific paragraphs with different temperatures or models → final light edit pass. This breaks up patterns.
- **Context & style memory.** Feed it examples of the user’s own previous human writing (if they have any) or target styles (e.g., “university business student, slightly casual but academic”).
- Test against multiple detectors regularly (Originality, GPTZero, Winston, etc.), not just one. Detectors keep updating, so what works today might slip in a month.

### Scaling technically:

Start simple. If it’s prompt-based right now, look into fine-tuning a smaller open model (Llama 3.1 8B or Mistral variants, or even Phi-3) on a dataset of human vs AI text pairs. Or use LoRA so it’s cheap.  
For speed/cost at scale, mixture-of-agents or routing between models (cheap one for first pass, stronger for refinement) works well. Monitor token usage obsessively.

### Deployment:

- **MVP:** FastAPI backend + simple React/Next.js frontend. Host on Vercel + Railway/Render/Fly.io. Super cheap to start.
- Add user accounts (Supabase or Clerk for auth), credit system (they buy “words” or monthly subs), and a queue if it gets heavy.
- Later: Turn it into a proper API so other devs or power users can integrate it. Chrome extension would be killer for students (direct “Humanize” button in Google Docs, etc.).

### For Design:

You can check these two sites which were shared by some users in 1h 🙂

[21st.dev](https://21st.dev/home)  
[getdesign.md](https://getdesign.md/)

---

<div class="post-metadata">

**Author:** ![Robert\_McClain](https://onehack.st/user_avatar/onehack.st/robert_mcclain/32/138460_2.png) [@Robert\_McClain](https://onehack.st/u/Robert_McClain)\
**Post date:** [May 9, 2026, 9:26pm UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/7 "2026-05-09T21:26:27Z")

</div>

Nice advise!

---

<div class="post-metadata">

**Author:** ![Margaret](https://onehack.st/user_avatar/onehack.st/margaret/32/174872_2.png) [@Margaret](https://onehack.st/u/Margaret)\
**Post date:** [July 6, 2026, 7:08am UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/8 "2026-07-06T07:08:49Z")

</div>

# 🛠 Build your own Grammarly. From free parts. $0.

The “AI writing app” they charge you $20/mo for is ~20 free open-source tools taped together — the model API is 5% of it. Here’s the full parts list. Runs on your machine, your text never leaves, no coding degree needed.

* * *

 ![image](https://onehack.st/uploads/default/original/3X/2/c/2ccf2e2f5b92589d10e8742d0210cf84d17b8827.jpeg)

* * *

Grab the shelves you want, ignore the rest. 👇

> **😮 where this actually pays off (real situations)**
>
> ```auto
> ┌──────────────────────────────┐
> │ ✍️ editor people type in │
> ├──────────────────────────────┤
> │ 🧠 grammar / style engines │
> │ 🎛️ clean output + PII scrub │
> │ 📚 memory (RAG + search) │
> │ 🔌 model gateway + local AI │
> ├──────────────────────────────┤
> │ 👀 watch · 💸 cache · 🏦 biz │
> └──────────────────────────────┘
> all free · all open-source
> 
> ```
> 
> - 💼 **Agency/freelancer:** hand each client a private “brand voice” editor instead of paying per-seat Grammarly Business — their drafts never leave your server.
> - 🌍 **You write in a language Grammarly ignores:** LanguageTool self-hosts 25+ languages, free.
> - 🚀 **Indie/student:** ship a niche writing tool (for lawyers, screenwriters, doctors) this weekend — fork an existing app, no funding, no degree.
> - 🔒 **Privacy work (legal/medical/journalism):** run the whole thing offline, text never touches OpenAI.
> - 💸 **Paying for 5 AI subs?** One gateway swaps them all and kills the per-token bill — one bill, not five.

> **🧠 the 90% that isn't a ChatGPT wrapper**
>
> Grammar/style checkers that run _under_ the AI, so you’re not paying a model to catch a typo.
> 
> | Tool | What |
> | --- | --- |
> | [**Harper**](https://github.com/Automattic/harper) · [**site**](https://writewithharper.com/) | Rust grammar checker, runs in-browser, private, instant |
> | [**LanguageTool**](https://github.com/languagetool-org/languagetool) · [**self-host**](https://www.codeslikeaduck.com/posts/selfhostlanguagetool/) | 25+ language grammar server, Docker |
> | [**Vale**](https://alternativeto.net/software/grammarly-grammar-checker/?license=opensource) | Enforce your own house style = your “brand voice” feature |
> | [**proselint**](https://github.com/amperser/proselint) · [**write-good**](https://github.com/amperser/proselint) · [**retext**](https://github.com/amperser/proselint) · [**alex**](https://github.com/topics/alex?o=desc&s=updated) | Cliché/weasel-word/passive/insensitive catchers |
> | [**textstat**](https://github.com/textstat/textstat) | Reading-level scores = your “readability” feature |
> | [**prose-linter MCP**](https://github.com/topics/prose?l=python) | 5 linters + AI-slop detector, pre-bundled. Fork it |
> | [**Grammarly-alt map**](https://www.localalternative.io/alternatives/grammarly) | Harper+Vale+local AI = Grammarly Premium, 100% local |

> **🔌 one plug for every model + run it free on your box**
>
> A gateway = one API for every model (swap OpenAI↔Claude↔free-local by changing one line, auto-failover). Local inference = run open models yourself, data stays home, per-token bill → $0.
> 
> | Tool | What |
> | --- | --- |
> | [**LiteLLM**](https://github.com/BerriAI/litellm) · [**setup**](https://medium.com/@adnanmasood/using-litellm-as-an-open-source-llm-proxy-the-llm-gateway-playbook-part-2-c50166ac1446) | One endpoint, 100+ providers, budgets, fallbacks |
> | [**OpenRouter**](https://openrouter.ai/blog/insights/llm-gateway/) · [**Portkey**](https://pinggy.io/blog/best_ai_llm_routers_openrouter_alternatives/) | 400+ models one key / gateway + cache + PII |
> | [**Helicone**](https://inworld.ai/resources/best-llm-gateways) · [**Kong**](https://contabo.com/blog/best-llm-gateways/) · [**Cloudflare**](https://contabo.com/blog/best-llm-gateways/) · [**TrueFoundry**](https://www.truefoundry.com/blog/litellm-vs-openrouter) · [**openziti**](https://blog.openziti.io/comparing-open-source-llm-gateways) | Router variants (obs / Kong / edge / K8s / zero-trust) |
> | [**Ollama**](https://github.com/ollama/ollama) | “Docker for LLMs” — one command runs a model locally |
> | [**vLLM**](https://github.com/vllm-project/vllm) · [**SGLang**](https://github.com/sgl-project/sglang) | Production serving, 3-8x throughput under load |
> | [**LocalAI**](https://github.com/mudler/LocalAI) · [**LM Studio**](https://lmstudio.ai) · [**Open WebUI**](https://github.com/open-webui/open-webui) | OpenAI-drop-in / GUI / self-host ChatGPT UI |
> | [**Olla**](https://thushan.github.io/olla/) | Puts every local engine behind one address |
> | [**Vercel AI SDK**](https://github.com/vercel/ai) · [**docs**](https://ai-sdk.dev/docs/introduction) · [**RAG template**](https://vercel.com/templates/next.js/ai-sdk-internal-knowledge-base) | Streaming chat + backend glue, clonable skeleton |
> | [**awesome-llm-services**](https://github.com/av/awesome-llm-services) | Master list, every self-host server |

> **🎛️ force clean data out + scrub the PII**
>
> Make the model return proper data (not a paragraph you regex), block junk input, strip names/cards _before_ they hit a log.
> 
> | Tool | What |
> | --- | --- |
> | [**BAML**](https://github.com/boundaryml/baml) | Prompt → typed, testable function |
> | [**Agenta**](https://agenta.ai) · [**Pezzo**](https://github.com/pezzolabs/pezzo) · [**Latitude**](https://github.com/latitude-dev/latitude-llm) · [**topic**](https://github.com/topics/prompt-management) | Git-style prompt versioning + rollback |
> | [**Instructor**](https://github.com/567-labs/instructor) | Define a shape → validated JSON every time, auto-retry |
> | [**Outlines**](https://github.com/dottxt-ai/outlines) | Forces valid output _during_ generation, no retries |
> | [**PydanticAI**](https://github.com/pydantic/pydantic-ai) · [**Guidance**](https://github.com/guidance-ai/guidance) · [**Marvin**](https://github.com/PrefectHQ/marvin) | Typed agents / constrained templating / simple syntax |
> | [**NeMo Guardrails**](https://github.com/NVIDIA-NeMo/Guardrails) · [**Superagent**](https://github.com/superagent-ai/ai-sdk) · [**AI SDK mw**](https://vercel.com/blog/ai-sdk-4-2) | Block jailbreaks + injection + PII, few lines |
> | [**Presidio**](https://github.com/microsoft/presidio) · [**example**](https://github.com/lotharschulz/pii-redaction-guard) | Find + mask names/cards/IDs (the PII standard) |

> **📚 give it a memory — RAG done cheap**
>
> RAG = let the AI answer from _your_ docs. Store meaning (vector DB), turn text into meaning (embeddings), pull source text in, add a search box.
> 
> | Tool | What |
> | --- | --- |
> | [**LlamaIndex**](https://github.com/run-llama/llama_index) · [**free build**](https://developers.llamaindex.ai/python/examples/low_level/oss_ingestion_retrieval/) | Best retrieval brain + full OSS walkthrough |
> | [**LangChain**](https://github.com/langchain-ai/langchain) · [**Haystack**](https://haystack.deepset.ai) · [**RAGFlow**](https://github.com/infiniflow/ragflow) | Chains / auditable pipelines / deep-PDF |
> | [**pgvector**](https://github.com/pgvector/pgvector) · [**scale**](https://awesomeagents.ai/tools/best-ai-rag-tools-2026/) · [**Qdrant**](https://qdrant.tech/documentation/frameworks/llama-index/) · [**Redis**](https://github.com/Danielskry/Awesome-RAG) | Vector search (pgvector = inside Postgres you already run) |
> | [**sentence-transformers**](https://github.com/huggingface/sentence-transformers) · [**BGE**](https://github.com/FlagOpen/FlagEmbedding) · [**reranker**](https://huggingface.co/BAAI/bge-reranker-v2-m3) · [**fastembed**](https://github.com/Anush008/fastembed-rs) | Text→coordinates + re-sort (reranker = cheapest accuracy jump) |
> | [**Firecrawl**](https://github.com/firecrawl/firecrawl) · [**Crawl4AI**](https://github.com/unclecode/crawl4ai) · [**MarkItDown**](https://github.com/microsoft/markitdown) | Web page / any file → clean Markdown |
> | [**Meilisearch**](https://github.com/meilisearch/meilisearch) · [**Typesense**](https://github.com/typesense/typesense) · [**Tantivy**](https://github.com/quickwit-oss/tantivy) · [**Quickwit**](https://github.com/quickwit-oss/quickwit) · [**Sonic**](https://github.com/valeriansaliou/sonic) | Instant typo-tolerant search box |
> | [**Awesome-RAG**](https://github.com/Danielskry/Awesome-RAG) · [**scraping list**](https://github.com/h4ckf0r0day/awesome-ai-web-scraping) | Master lists |

> **✍️ the thing people actually type in**
>
> Rich-text editor + autosave + two people typing at once (that fight-prevention magic = a CRDT) + rendering AI output as it streams without glitching.
> 
> | Tool | What |
> | --- | --- |
> | [**Tiptap**](https://github.com/ueberdosis/tiptap) · [**site**](https://tiptap.dev/product/editor) | Headless editor (you own the look), the default |
> | [**ProseMirror**](https://github.com/ueberdosis/tiptap) · [**Lexical**](https://medium.com/@faisalmujtaba/tiptap-vs-lexical-which-rich-text-editor-should-you-pick-for-your-next-project-17a1817efcd9) | The engine / Meta’s, built for scale |
> | [**BlockNote/Plate/Slate/Quill**](https://github.com/topics/rich-text-editor) · [**BlockSuite**](https://github.com/topics/rich-text-editor) | Notion-style blocks + the rest |
> | [**Yjs**](https://github.com/ueberdosis/tiptap) | The CRDT: autosave, offline, undo/redo, live cursors |
> | [**Hocuspocus**](https://github.com/ueberdosis/tiptap) · [**SyncedStore**](https://syncedstore.org/docs/advanced/richtext/) · [**Automerge**](https://gist.github.com/0xdevalias/2fc3d66875dcc76d5408ce324824deab) · [**Liveblocks**](https://liveblocks.io/blog/which-rich-text-editor-framework-should-you-choose-in-2025) · [**Etherpad**](https://gist.github.com/0xdevalias/2fc3d66875dcc76d5408ce324824deab) | Ready-made sync servers |
> | [**Streamdown**](https://github.com/vercel/streamdown) · [**markstream-vue**](https://github.com/Simon-He95/markstream-vue) · [**RN**](https://github.com/software-mansion-labs/react-native-streamdown) | Render half-finished AI Markdown mid-stream |

> **👀 x-ray every AI call before users hit the bug**
>
> See what it’s doing (cost, speed, what broke) + unit-test writing quality so a prompt tweak doesn’t quietly ruin everything.
> 
> | Tool | What |
> | --- | --- |
> | [**Langfuse**](https://github.com/langfuse/langfuse) · [**site**](https://langfuse.com/) | Self-host: tracing + prompts + evals. The default |
> | [**Phoenix**](https://openobserve.ai/blog/llm-observability-tools/) · [**Helicone**](https://machinelearningmastery.com/llm-observability-tools-for-reliable-ai-applications/) | RAG-debug / swap-one-URL logging |
> | [**OpenLLMetry**](https://openobserve.ai/blog/llm-observability-tools/) · [**Lunary**](https://openobserve.ai/blog/llm-observability-tools/) · [**Opik**](https://www.firecrawl.dev/blog/best-llm-observability-tools) · [**Weave**](https://www.firecrawl.dev/blog/best-llm-observability-tools) · [**Evidently**](https://www.firecrawl.dev/blog/best-llm-observability-tools) · [**OpenObserve**](https://openobserve.ai/blog/llm-observability-tools/) | OTEL / lightweight / drift / one-backend |
> | [**promptfoo**](https://github.com/promptfoo/promptfoo) | Compare prompts side by side + CI gates |
> | [**DeepEval**](https://github.com/confident-ai/deepeval) · [**RAGAS**](https://github.com/explodinggradients/ragas) · [**guide**](https://github.com/hparreao/Awesome-AI-Evaluation-Guide) | “pytest for LLMs” / RAG scoring / starter |

> **💸 keep it fast and dirt cheap**
>
> Move slow work off the request (queue), reuse repeat answers (cache), count tokens before you’re billed.
> 
> | Tool | What |
> | --- | --- |
> | [**BullMQ**](https://bullmq.io/) · [**rate-limit**](https://docs.bullmq.io/guide/rate-limiting) | Job queue: retries, cron + per-customer limits |
> | [**Trigger.dev**](https://trigger.dev/) · [**Inngest**](https://www.inngest.com/) | Long AI jobs / event-driven, no Redis |
> | [**Graphile Worker**](https://github.com/graphile/worker) · [**pg-boss**](https://github.com/timgit/pg-boss) | Use existing Postgres as the queue |
> | [**GPTCache**](https://github.com/zilliztech/GPTCache) | Semantic cache — cuts API cost ~10x |
> | [**tiktoken**](https://github.com/openai/tiktoken) · [**go**](https://github.com/pkoukk/tiktoken-go) · [**rs**](https://github.com/zurawiki/tiktoken-rs) | Count tokens locally, never blow the context window |
> | [**tokencost**](https://github.com/AgentOps-AI/tokencost) | One call → $ cost across 400+ models |

> **🏦 the boring bits that make it a business**
>
> Login, subscriptions, analytics, onboarding email, hosting. Auth alone = ~40-60 hrs — don’t build it, fork it.
> 
> | Tool | What |
> | --- | --- |
> | [**Open SaaS**](https://opensaas.sh/) | Free boilerplate: auth + Stripe/Lemon/Polar + admin |
> | [**Supabase**](https://supabase.com/) | Postgres + auth + storage + realtime + pgvector |
> | [**Basejump**](https://usebasejump.com/) · [**MakerKit**](https://makerkit.dev/supabase) · [**Supastarter**](https://supastarter.dev/) · [**SvelteKit**](https://saasstarter.work/) · [**Next+Stripe**](https://github.com/umesh00/next-saas-supabase-stripe-lemon-starter) · [**Saas-Kit**](https://github.com/Saas-Starter-Kit/Saas-Kit-supabase) | Teams, roles, billing — forkable |
> | [**PostHog**](https://github.com/PostHog/posthog) | Analytics + funnels + flags + A/B + errors + replay, one install |
> | [**GrowthBook**](https://github.com/growthbook/growthbook) · [**OpenReplay**](https://github.com/openreplay/openreplay) · [**Plausible**](https://github.com/plausible/analytics) · [**Umami**](https://github.com/umami-software/umami) | Flags / replay / cookieless analytics |
> | [**Novu**](https://github.com/novuhq/novu) · [**React Email**](https://github.com/resend/react-email) · [**Listmonk**](https://github.com/knadh/listmonk) | Notifications / email in React / newsletters |
> | [**Coolify**](https://github.com/coollabsio/coolify) | Self-host Vercel: git-push + auto-SSL on a $5 box |
> | [**Dokku**](https://github.com/dokku/dokku) · [**Kamal**](https://github.com/basecamp/kamal) · [**CapRover**](https://github.com/caprover/caprover) · [**Dokploy**](https://github.com/Dokploy/dokploy) · [**Caddy**](https://github.com/caddyserver/caddy) | Lighter deploy + auto-HTTPS proxy |
> | [**Uptime Kuma**](https://github.com/louislam/uptime-kuma) · [**OneUptime**](https://oneuptime.com/) | Uptime + status page, one command |

> **🦾 bonus toys + the maps to all of it**
>
> | Tool | What |
> | --- | --- |
> | [**LangGraph**](https://github.com/langchain-ai/langgraph) · [**CrewAI**](https://github.com/crewAIInc/crewAI) · [**AutoGen**](https://github.com/microsoft/autogen) · [**smolagents**](https://github.com/huggingface/smolagents) · [**Mastra**](https://github.com/mastra-ai/mastra) · [**Agno**](https://github.com/agno-agi/agno) · [**roundup**](https://www.firecrawl.dev/blog/best-open-source-agent-frameworks) | Run draft→critique→revise for you |
> | [**Whisper**](https://github.com/openai/whisper) · [**faster**](https://github.com/SYSTRAN/faster-whisper) · [**speaches**](https://github.com/speaches-ai/speaches) · [**Coqui**](https://github.com/coqui-ai/TTS) · [**WhisperSpeech**](https://github.com/WhisperSpeech/WhisperSpeech) | Dictation in + read-aloud out (speaches = “Ollama for voice”) |
> | [**Reor**](https://github.com/reorproject/reor) | Local AI note app — editor + RAG in one repo, the reference build |
> | [**StoryCraftr**](https://github.com/raestrada/storycraftr) · [**app topic**](https://github.com/topics/ai-writing-assistant) | Long-form AI writer + forkable apps feed |
> | [**awesome-llm-apps**](https://github.com/Shubhamsaboo/awesome-llm-apps) · [**Awesome-LLMOps**](https://github.com/InftyAI/Awesome-LLMOps) · [**awesome-ai**](https://github.com/openbestof/awesome-ai) | 100+ fork templates + master indexes |

> **🎯 where do I even start?**
>
> | You want… | Grab |
> | --- | --- |
> | Cheapest possible start | Ollama + LiteLLM + pgvector + Tiptap — all local, $0 |
> | To not sound like AI slop | Harper + Vale + proselint under any model |
> | Total privacy, nothing leaves your box | Local AI + Presidio + self-host everything |
> | Ship a business this month | Open SaaS → PostHog → Coolify |
> | Just read one working app | Reor — the whole thing in one repo |

**The whole stack, free:** ingest → write → check → cache → ship. Every layer above is one folder of GitHub links.

* * *

 ![image](https://onehack.st/uploads/default/original/3X/e/1/e16b3fb6fd40d6e8e3ef925077d40faa937da115.jpeg)

* * *

⚡ It was never a $20 product. Just $0 parts nobody stacked in front of you — until now.

---

<div class="post-metadata">

**Author:** ![system](https://onehack.st/user_avatar/onehack.st/system/32/165705_2.png) [@system](https://onehack.st/u/system)\
**Post date:** [July 6, 2026, 7:09am UTC](https://onehack.st/t/built-my-own-ai-humanizer-tool-after-facing-issues-need-advice-on-scaling-launching/320851/9 "2026-07-06T07:09:23Z")

</div>

> [@Margaret](#):
>
> 🛠 Build your own Grammarly. From free parts. $0. The “AI writing app” they charge you $20/mo for is ~20 free open-source tools taped…

**♻ The Core-Community team’s AI turned the answer above into a better version.**