Turn your documents into an assistant that cites its sources.

Upload the policies, FAQs, and notes you already have. Ask a question in plain language and get an answer with the passages it came from.

Answers are grounded in your documents, not the open web.

Refund Policy.pdf
Ask AI
Example

What is the refund policy?

Customers can request a full refund within 30 days of delivery, as long as the item is unused and in its original packaging. Requests are reviewed within two business days, and approved refunds go back to the original payment method.

Sources

  • Refund Policy.pdf

    Customers may request a full refund within 30 days of delivery, provided the item is unused and in its original packaging.

  • Support Handbook.md

    Refund requests are reviewed within two business days. Approved refunds return to the original payment method.

An example answer about a refund policy, shown with the two source passages it was drawn from.

The problem

The knowledge is written down. It just isn't answerable.

You already have the documents you need. What's missing is a way to ask them a question and get back an answer you can check.

The answer exists, but nobody can find it

The policy lives in a PDF, the exception lives in a note somewhere else, and neither is where you looked first.

General AI tools invent the details

A model trained on the open web will confidently describe a refund window you never offered.

Nobody trusts an answer they can't check

Without a link back to the source passage, every reply still has to be verified by hand.

How it works

Three steps from documents to answers

No embedding configuration, no pipeline babysitting, and no vocabulary you need a machine learning background to follow.

  1. 1

    Create your workspace

    Set up a workspace — for yourself or for your whole team — then add a knowledge space for each area you want the assistant to cover.

  2. 2

    Upload your documents

    Drop in PDFs, FAQs, policies, or product docs. One action prepares them for chat; you never touch a processing pipeline.

  3. 3

    Ask questions and get cited answers

    Ask in plain language. Every answer arrives with the source passages it came from, so you can check the reasoning.

Features

Built around one job: answering from your own content

Document-based answers

Replies are generated from the documents you uploaded, so the assistant stays inside what you actually published.

Sources from your knowledge

Every answer lists the passages behind it, with the document each one came from.

Multi-workspace support

Keep separate workspaces for different clients, products, or teams, each with its own knowledge spaces.

Conversation history

Past chats are saved with their sources, so you can return to an answer and see what backed it.

Simple document processing

One action takes a file from uploaded to ready. Progress is visible the whole way through.

Workspace-scoped data

Documents, chats, and search are scoped to the workspace that owns them, enforced by the API on every request.

Use cases

Wherever the answer is already written down

Customer support

Give yourself, or a whole team of agents, a fast way to check refund windows, shipping rules, and account policies without opening four documents.

“How long does a customer have to return an opened item?”

Product documentation

Let the team ask the docs directly instead of scrolling through release notes and configuration guides.

“Which plans include the audit log?”

Internal team knowledge

Turn onboarding guides and process docs into something a new hire can question in plain language.

“What is the approval process for a new vendor?”

Policies and FAQs

Answer the same client question the same way every time, straight from the terms you actually published.

“What are my revision limits on a fixed-price project?”

Point it at your documents and ask a question.

Create a workspace, add a document, and see a cited answer in a few minutes.