The research is all there. Nobody can ask it a question.

CorpusKit is an open source research portal that runs on the Progress Agentic RAG platform. Load your documents into a knowledge box (a private, searchable collection) and CorpusKit turns it into a portal your researchers can question. Answers cite the page they came from.

On CorpusKit Cloud, Noice runs it for you.

Request a 14-day trial Try the demo

The demo is the CorpusKit documentation portal, which runs on CorpusKit.

Trials run for 14 days at yourname.corpuskit.cloud, by invitation.

Already have an account? Sign in

Data residency

Documents, their index and the knowledge graph are stored in Progress's Australian region (AWS Sydney). Other regions on request.

Where everything is stored

Who it is for

Organisations with years of research behind them: rural research and development corporations, government agencies, research institutes and industry bodies.

The reports are published, and most were funded so that people would use them. Finding the paragraph that answers a question still takes someone who knows where to look.

Researchers want an answer they can cite. Whoever runs the knowledge program wants something they can load and explain to IT.

For researchers

  1. What can I ask, and what do I get back?

    Ask a question in plain language and get an answer written from your organisation's own reports. Numbered citations open each report at the page used, so you can check any statement.

  2. How do I find things without knowing what to ask?

    Start with Explore, which suggests questions and topics from your own collection. Search by meaning or exact words, and filter by topic and document type. On a Pro licence, the library gives every document a readable title, summary and key takeaways, and holds audio and video too.

  3. Can I keep a question open for months?

    Yes. An investigation collects evidence and notes as you find them, over weeks or months, and exports to Word. The trail is still there when someone asks how a conclusion was reached.

  4. Can I see how topics connect?

    Yes. The knowledge graph shows the names, places and topics in your research and how they connect, with the passage behind each connection.

  5. Can it turn the research into a briefing?

    Yes. Generate a briefing, comparison, timeline or FAQ from the collection, each with numbered references. Knowledge checks set a short quiz on a topic at three levels, each question linked to its source.

  6. Can I trust the answer?

    Figures, populations, named studies and years are tested against the cited text, sentence by sentence. A sentence that fails is removed, and a note says so. Each answer is rated high, moderate or low confidence. See an answer checked below.

Portal features

  • Explore

    Suggested questions and topic rows from your own collection, so first-time visitors know what they can ask.

  • Search

    Match by meaning, exact words or both, best results first. Filter by topic and document type. A saved search re-runs daily and flags anything new.

  • Ask

    Conversations with numbered citations. The answer appears as it is drafted, the check runs in view, and the checked version then replaces it.

  • Library

    Every document with a readable title, summary and key takeaways written from its contents (on a Pro licence). PDFs, Office files and web pages, and on Pro, audio and video.

  • Knowledge graph

    The names, places and topics in your research and how they connect, with the passage behind each connection.

  • Investigations

    Long-running questions that gather evidence, notes and verdicts, and export to Word.

  • Generate

    Briefings, comparisons, timelines, FAQs, and pros-and-cons lists, each with numbered references. Sections the collection cannot support are left out and named.

  • Knowledge checks

    Short quizzes on a topic at three levels, each question linked to its source.

  • Manage

    Uploads, scheduled website re-syncs, label sets, automatic summaries and labels, and a check that flags any document to reload.

  • Access and accessibility

    Six user roles, from reader to administrator. Portals can be public, staff-only or restricted, with an exportable audit log. Designed for WCAG AA contrast and full keyboard use.

Hosting

Noice hosts the portal at your own web address, such as research.yourorg.au, or at yourname.corpuskit.cloud. We apply upgrades and handle support in Melbourne business hours.

To set up a portal, CorpusKit drafts its topics, suggested questions and labels from your documents, and our team or yours checks the draft.

Sign in with an emailed link, or with your organisation's single sign-on (Microsoft Entra ID or another SSO provider), set up on request.

Hosted

A$750 a month

  • Your own web address, or yourname.corpuskit.cloud
  • Upgrades and security fixes included
  • Your team looks after the collection, in the portal
  • A monthly usage report, and limits you set
  • Email support in Melbourne business hours

Managed

A$1,750 a month

  • As Hosted, and Noice also loads new material each month
  • Noice looks after the collection, up to 8 hours a month in all: document loads, tuning, a quality review and changes
  • Email support as Hosted, and a quarterly review

Prices exclude GST and the Progress Agentic RAG licence. Compare plans and prices

Custom domains

On Hosted and Managed, your portal can live at your own web address, such as research.yourorg.au, or at yourname.corpuskit.cloud.

Trials use yourname.corpuskit.cloud. A portal on your own web address keeps that address if you leave.

  • https://research.yourorg.auHosted and Managed
  • https://yourname.corpuskit.cloudTrials, Hosted and Managed

Data storage

Documents, index and knowledge graph

Your documents, index and knowledge graph are held in your knowledge box. Noice creates new boxes in Progress's Australian region (AWS Sydney).

Sign-in emails, settings and sessions

On CorpusKit Cloud, sign-in emails, portal settings, sessions and saved investigations are held by our hosting provider, configured for its Oceania region.

Language models

Language models receive document text for summaries and labels, and each question with the passages found for it. Available models vary by Progress plan; ask us where one runs before you load documents.

Your document system

Your document system stays. CorpusKit reads a copy of your documents. The originals stay where they are.

Security

Open source

The code is released under Apache 2.0. You can change it and run it yourself, or have Noice run it.

The code is public at github.com/noicework/corpuskit. It runs on common cloud platforms or as a container, with our help if you want it. On the reference deployment, staff sign in with Microsoft Entra.

Portability

The licence and knowledge box are in your organisation's name, so your documents, index, labels and graph stay with you. Each investigation and saved session exports to Word. You can run the open source portal yourself, and a portal on your own web address keeps that address.

About Noice

Noice is a Melbourne digital agency and Progress partner. We build, host and support CorpusKit. Noice is on the Victorian Government eServices Register and holds professional indemnity, public liability and cyber insurance.

Getting started

  1. Ask the demo

    Ask the documentation portal a question at demo.corpuskit.org.

    Try the demo

  2. Ask for a trial

    Ask for a 14-day trial, using reports you have already published. Trial documents are held in Noice's Progress account in the Australian region. About 30 days after the trial starts we delete the trial knowledge box and suspend the portal, unless you go ahead.

    Request a 14-day trial

  3. Book a demo

    Email us to book a demo on your own collection.

    hello@noice.net.au