DocuTrain takes a pile of documents — PDFs, PowerPoint decks, Google Slides, training manuals, clinical guidelines — and turns each one into its own focused AI chatbot. Users ask questions in plain language; the bot answers with grounded citations back to the source pages and slides. Nothing more, nothing less. No hallucinated facts, no wandering off-topic.

That's the elevator pitch. But the surface area underneath is wide, and most of the platform's power comes from two things you configure: documents and owner groups. This blog series walks through both.

The two things you'll configure

DocuTrain is structured around a clean hierarchy:

Owner (your organization or team)
  ├── Branding, plan, custom domain, chat-assistant persona
  ├── Document 1
  │     └── Access level, AI prompt, UI options, attachments, quiz, abstract...
  ├── Document 2
  ├── Collection (a bundle of documents with shared access)
  └── ...

Most of what makes DocuTrain feel like your product — not a generic chatbot — lives in those two configuration surfaces. The rest of this series unpacks them.

What ships out of the box

Before we get into configuration, here's what every DocuTrain deployment includes:

  • Rich document ingestion. Upload PDFs, PowerPoint (.pptx), Google Slides, or paste text. DocuTrain extracts text, captures slide images, generates draft titles and descriptions for slides, and prepares everything so questions can be answered accurately.
  • Answers tied to your sources. Responses cite where the information came from — page numbers and slide thumbnails when available — instead of guessing from thin air.
  • Reliable AI answers. DocuTrain handles the AI stack for you and keeps every answer grounded in your uploaded content — you focus on the document and the prompt, not on provider details.
  • Separate workspaces by design. Every document and conversation belongs to an organization in DocuTrain, so one customer's library is never mixed up with another's.
  • Granular access control. Public, registered-users-only, passcode-protected, owner-group-restricted, or admin-only — set per document or per collection.
  • Per-document branding and UI. Welcome messages, suggested questions, downloads sidebar, references, image galleries, quizzes, and embedded widgets all toggle on and off per document.
  • Per-owner branding. Logo, accent color, default cover image, landing page layout, custom domain, chat-assistant persona — all configurable.
  • Integrations. Embed as a widget, connect Telegram bots, send transactional email through Mailgun.
  • Plan tiers and quotas. Free, Pro, Business, and Unlimited tiers control how many documents you can publish and which features you can use.

The blog series

This series is broken into six posts. Read them in order if you're new; jump around if you know what you're looking for.

  1. Configuring a Document End-to-End — A walk through every tab of the document editor: basic info, content, UI options, messages, attachments, embed, telegram, and admin metadata.
  2. Access Control: Public, Registered, Passcode, and Beyond — The five access levels, when to use each, and how collections layer their own access tokens on top.
  3. Steering the AI: Prompts, Safety, Abstracts, and Quizzes — Per-document custom prompts, the safety reviewer, abstract generation, and quiz generation.
  4. Owner Groups: Branding, Landing Pages, and the Chat Assistant — How owner groups work: logos, accent colors, landing-page layouts, the chat-assistant persona, fun facts, and medication-warning settings.
  5. Collections, Categories, and Organizing Many Documents — Document collections with shared access tokens, two-level category hierarchies, language associations, and discoverability.
  6. Plans, Quotas, and Integrations — The plan-tier system, what each tier unlocks, custom domains, embed widgets, Telegram bots, and email.

Let's start with the document editor.