DEV Community

Mārtiņš Veiss
Mārtiņš Veiss

Posted on

Why We Built AutoBot: The WordPress of AI

Three years ago I started building AutoBot because I was tired of renting my own intelligence.

Every AI tool I used followed the same playbook: send your data to our servers, pay monthly, accept our terms, trust us not to read your prompts. The model was the product. You were the user. Your data was the inventory.

I wanted something different. I wanted an AI that felt like mine.


Your data. Your AI.

That's the line we built everything around.

WordPress gave everyone a website. Before WordPress, having a web presence meant renting space on someone else's platform, playing by their rules, losing your content if they shut down. WordPress flipped that. You install it, you own it, you extend it, you run it forever.

AutoBot is that for AI.

Self-hosted. Open source. Yours to extend. The AI platform that belongs to you — not to us.


Feed it your world

The core feature isn't the chat interface. It's the knowledge base.

Drop in your docs. Upload your codebase. Paste your business processes. AutoBot's RAG engine turns your raw files into a searchable, queryable AI layer that actually knows your domain — not just what some model was trained on two years ago.

Feed it your docs. Your codebase. Your business.
It learns what you know. It stays where you are.

This is the part that changes things. An AI that knows your codebase gives better answers than a generic model. An AI that's read your legal documents is more useful than one guessing at your jurisdiction. An AI trained on your patient intake forms is more reliable than one pattern-matching across the internet.

An AI that's actually about you.


Pick your brain

AutoBot is not opinionated about which LLM powers it. That's your call.

Want to run fully local? Plug in Ollama. LM Studio. llama.cpp. Anything with an OpenAI-compatible endpoint.

Prefer GPT-4 or Claude for the heavy lifting? Connect your API key. Your data stays on your machine — your prompts go to the model, but your knowledge base documents don't.

That distinction matters. The brain phones home. Your documents don't.


You shouldn't have to ask permission

The thing that finally broke me on cloud AI wasn't the pricing. It was the 2 AM email.

"We're updating our terms of service effective next month. By continuing to use the service you agree to..."

Your data never leaves your machine — whatever brain powers it.
No rate limits on your knowledge. No vendor changing the rules on you overnight.

Deploy once. Run it your way. Forever.


Most AI is rented. AutoBot is yours.

You decide what it does. You decide where it runs. You decide who sees it.

No subscription. No surveillance. No one reading your prompts.
Install it. Own it. Run it forever.

For developers who've been burned by API deprecations, pricing pivots, and terms changes — this is for you.

For law firms, medical startups, and anyone in a regulated industry — your data never touches our servers. No cloud vendor to breach. No third party holding your keys. What stays on your machine stays yours — full stop. You control the perimeter. We just give you the tools.


Get started in 5 minutes

git clone https://github.com/mrveiss/AutoBot-AI.git
cd AutoBot-AI
docker compose up -d
Enter fullscreen mode Exit fullscreen mode

Open http://localhost:3000. Connect your LLM. Feed your first document.

That's it. You're running your own AI.


AutoBot is open source. If this resonates, star us on GitHub, join the community discussions, or sponsor the project.

Top comments (0)