1
1 Comment

I built an AI chatbot SaaS in 2 months that lets anyone add a document-trained chatbot to their website with one line of code

Six months ago I had a problem. I was helping a friend's small business answer the same customer questions over and over — "What's your return policy?", "Do you ship internationally?", "How do I use this product?"
The answers were all in their docs. But no one was reading them.
That's when the idea hit me: what if the docs could just… answer back?

🛠️ The Build (2 months, nights & weekends)
I'm a solo developer, and I gave myself a hard constraint: ship something real in 2 months or kill the idea.
Here's what I actually built:

A dashboard where you upload your documents (PDF, DOCX, TXT, CSV)
The system runs RAG (Retrieval-Augmented Generation) — it chunks your docs, indexes them, and uses them as context for GPT responses
A lightweight embeddable widget (under 8KB gzipped, Shadow DOM isolated, fully responsive)
One script tag to drop it on any site — HTML, WordPress, Shopify, Wix, Squarespace, React, Vue, Webflow — all covered

No fluff. No bloat. Just: upload docs → get a chatbot → embed it anywhere.

🚀 What it looks like in practice
You paste this into your website:
html<script
src="https://documind.website/widget.js"
data-chatbot-id="your_embed_code"
defer

</script>
That's it. A floating chat bubble appears. Your visitors ask questions. The AI answers from your documents — not from the general internet, not hallucinated facts. Your content.
You can customize brand colors, welcome messages, position, and avatar from the dashboard. No code needed.

📈 Who this is for
I built this for:

SaaS companies tired of repetitive support tickets
E-commerce stores that want to answer product questions 24/7
Agencies who want to white-label a chatbot for their clients
Content creators & educators who want their audience to explore their material interactively
Startups who can't afford a full support team yet

If you have documents and a website, you can have an AI support agent running today.

🔗 Live Links

🌐 Product: https://documind.website
📚 Docs & Integration Guide: https://documind.website/docs

💬 Honest reflection
Two months is fast. There are things I want to improve — analytics, more file format support, team collaboration features. But I'd rather have something real in people's hands than a perfect product that never ships.
If you're a founder, developer, or business owner dealing with repetitive questions from customers or users — I'd genuinely love for you to try it and tell me what's broken.
What's the most annoying question your customers keep asking? Drop it below — I'm curious.

Email: [email protected]

on March 11, 2026
  1. 1

    Love this approach — grounding strictly to uploaded docs to cut hallucinations is huge for reliable customer support without risking wrong info. I've run into the same need for control. With SupportBridge, I'm going even stricter: AI replies verbatim from approved FAQs only, max one auto-reply per ticket, and auto-escalates on anything sensitive like refunds or passwords. How are you planning to handle escalation or risky queries that go beyond the docs?

Trending on Indie Hackers
Your build-in-public audience is not your market. I learned the difference the slow way. User Avatar 202 comments I built a WhatsApp AI bot for doctors in Peru — launched 3 weeks ago, 0 paying customers, and stuck waiting for Meta to approve my app User Avatar 62 comments Built a "stocks as football cards" thing. 5 days in, my launch tweet got 7 views. What am I missing? User Avatar 33 comments From broke and burned out as a PM, to launching my SaaS and optimizing my health User Avatar 32 comments Why Claude Skills Are Becoming Important for Tech Careers User Avatar 24 comments I kept starting projects and dropping them. So I built a system that wouldn’t let me User Avatar 23 comments