Fiverr
Hire us for a guided setup, a repair, or a website help chat on a Fiverr order.
Open FiverrFor Shops, Salons and Homestays · 1.1.2
Docs Answer Bot
Answers customer questions from your own files, and shows where the answer came from. Optional AI answers are off by default.
Sound familiar?
- Customers ask the same questions late at night and nobody replies.
- Prices, room rates and policies are scattered across Word files, sheets and PDFs.
- New staff answer slowly, or give the wrong price.
What this shows
- Answered from README.md
- Extractive mode, not AI
- Python 3.10+
How we tested it
We asked it questions in a private Telegram chat. It answered from the product README and showed the source. This is version 1.1.2, which quotes at most two short excerpts.
| Check | Result |
|---|---|
| Version | 1.1.2 |
| Tests | 82 tests passed on Linux and on Windows, plus a clean install and a quick check after that. |
| Linux | PASS |
| Windows | PASS |
| Docker | Not in this package |
| SHA-256 of the download | aef36c8c1fe6c13a448788966976b8104bce2d9b42eb176d89f981dc2fc50f47 |
What you get
- Load your own files: Word, PDF with text, Markdown, CSV and plain text
- Answers quote your files and show the source, so you can check them
- A clear "I don't know from your documents" reply when nothing matches
- Ask in a local web page, or through an optional Telegram bot for a private chat
- Looking through your files stays on your computer
- Optional AI answers are off by default; if you turn them on, that text is sent to the AI service you set up
What You Need to Run This Tool
- A Windows or Linux computer or VPS you control — we don't host it for you
- Python 3.10 or later
- Your own documents: Word, text-based PDF, Markdown, CSV or plain text. Scanned PDFs need text recognition first
- Poppler's
pdftotexton your PATH if you want to load PDFs - Optional: a Telegram bot token and approved chat IDs, if you use the Telegram path
- Optional generated answers are off by default. To turn them on you set
LLM_BASE_URL,LLM_API_KEYandLLM_MODELto an endpoint you own or pay for. Question and passage text then leave your machine for that endpoint, and that provider's fees apply
Technical details
Looking through your files and pulling out matching passages stays on your computer. Optional AI answers are off by default. If you turn them on, you point the tool at an AI service you set up (LLM_BASE_URL, LLM_API_KEY and LLM_MODEL); the text of those questions and passages is sent to that service, and that service may charge you. The tool needs Python 3.10 or later, and Poppler's pdftotext for PDFs.
How it works
- Install the package in a Python 3.10+ environment on a computer you control.
- Start the local web app and add your own documents.
- Ask a question. The app quotes the passages it used, with file and page.
- Optionally run the native Telegram bot against a private, approved chat, using a bot token you provide.
How to set it up
- Unzip the download package and create a Python virtual environment.
- Install requirements.txt. PDF text extraction needs Poppler’s pdftotext on your PATH.
- Start the local web server and add your documents.
- Ask a question in the app. For Telegram, set your bot token and approved chat in the process environment, then run telegram_bot.py.
Packages
Download it yourself for $19 USD on Gumroad, or book setup help on Fiverr or Upwork.
Self-serve
$19 USD
download package, README, and a 7-day bug-fix window through Gumroad messages.
Basic
$39 USD
Local web app, Answers quoted from your files with citations, README, 7-day bug-fix support via Fiverr or Upwork messages.
Standard
$79 USD
Basic plus native Telegram bot setup guidance in a private test chat, and a short walkthrough.
Premium
$129 USD
Standard plus help wiring your own compatible model endpoint (you provide the key) and a 14-day support window.
Support stays on the platform you purchased through. Requirements: Python 3.10+ on Windows or Linux. You provide your own bot token when a Telegram connection is needed. Sepxo Labs does not receive your tokens or documents.