Skip to content

Analyse your documents

Analyze documents with AI and get answers that point to the page

Upload PDFs, Word files and decks, then summarize, ask questions, extract tables or compare documents. Answers from a PDF point back to the page they came from, and outside facts are researched and cited.

Jobs to try

  • One-page PDF summary: Summarize the attached PDF in one page for a busy executive: main argument, key numbers, risks and open questions, with page references for every point.
  • Contract terms table: Read the attached contract and list every obligation, deadline, renewal term, termination right and liability cap in a table, with the clause and page for each.
  • Extract tables to Excel: Extract every financial table from the attached PDF into an Excel workbook, one sheet per table, with the page number each table came from.
  • Research paper briefing: Read the attached research paper and explain the question, method, sample, main findings and weaknesses in plain English, with page references. Then find newer studies on the same question and cite them.
  • RFP response checklist: Go through the attached RFP and build a checklist of every requirement, deadline and evaluation criterion, with page references, as an Excel file.
  • Report to board deck: Turn the attached 80-page industry report into an 8-slide editable PowerPoint for our board, with the key figures as native charts and page references in the speaker notes.

What you get back

  • Answers with page references: Summaries and answers in chat where each point points back to the page of the PDF it came from, so you can open the passage and check it.
  • Extraction table (.xlsx): Fields, terms, figures and tables pulled out of your documents into an Excel workbook, one row per item, with the source page beside each value.
  • Briefing memo (.docx or .pdf): A written summary or comparison with real headings, built from your documents and any web research you asked for, with sources kept inside.
  • Summary deck (.pptx): An editable PowerPoint built from the document, with native charts, in one of six themes, and references in the speaker notes.

Researched and cited by default

No search toggle and no research mode to switch on. Rixx researches the live web before it answers a factual question, reads the files you give it, and cites every fact it uses.

Reading documents is slow, and skimming them is risky. Rixx reads the whole file, answers your question and shows you where in the document each answer came from. For PDFs, that means a page reference you can open and check in seconds.

Rixx is an AI agent, so it does more than answer. It can extract every table into Excel, line up five contracts side by side, check a report's figures against the live web and turn the result into a memo or a deck.

Four jobs Rixx does with documents

1. Summarize. One page, one paragraph or a detailed section-by-section summary. Say who it is for, and Rixx writes for that reader: a board summary leads with decisions, study notes lead with concepts.

2. Ask. Chat with your PDF the way you would with a colleague who has read it. "What does the contract say about early termination?" "Which assumptions drive the forecast on page 40?" Answers carry a page reference.

3. Extract. Pull fields, terms, numbers and tables into a structured table. Rixx runs Python and other tools in its own private workspace for the tables and figures inside documents, so totals are calculated rather than copied.

4. Compare. Line up several documents to find where they agree, where they differ and what each one leaves out. See summarize multiple PDFs for a step-by-step guide.

An AI PDF summarizer that cites the page

A summary leaves things out by design, so you need to know where each point came from. When Rixx answers from a PDF, its points carry references like "p. 14", linked to the passage. Open the ones that matter before you rely on them.

When you ask Rixx to bring in outside context, such as whether a market size quoted in a 2023 report has since been revised, it researches the live web and cites those sources inline. Document references and web citations sit side by side, so you can always tell which is which.

Good prompts for summaries:

  • "Summarize this in five bullet points for a CFO, with page references."
  • "Give me the argument of each chapter in two sentences."
  • "List every number in the executive summary and the page it is supported on."
  • "What does this document assume without evidence?"

Chat with a PDF, or a hundred

Ask follow-up questions in the same chat, and Rixx keeps the context. Add more files as the work grows: up to 5 files per upload, 50 MB each, and up to 100 documents in one chat. PDFs can run to 1,000 pages.

For documents you return to often, such as a data room, a set of policies or a reading list, keep the files, chats and instructions together in a Project.

Use cases

Document Typical job Output
Annual reports and filings Key financials, guidance changes, risk factors across years Excel table plus a one-page memo
Contracts and agreements Obligations, deadlines, renewal and liability terms Clause table with page references
Research papers Method, sample, findings and weaknesses, plus newer studies Briefing note with citations
RFPs and tenders Requirement checklist and evaluation criteria Excel checklist
Industry and analyst reports Market figures, checked against current sources Summary deck with native charts
Policies and manuals Answers to specific questions, with the exact section Answers in chat

Analyze documents with AI, then check them against the web

Documents age. A market report from two years ago may quote a figure that has since been revised, and a competitor named in a pitch deck may have changed its pricing. Rixx can read your document and research the live web in the same job:

  • "Check every market-size figure in this report against the latest published sources."
  • "This deck names six competitors. Find each one's current pricing and cite it."
  • "Which claims in this white paper are supported by independent research?"

The result separates what the document says from what current sources say, with both cited.

From answer to file

Once you have what you need, ask for a file. A job builds it in the background, renders it, reviews it and fixes problems before handing it back:

  • An Excel workbook of extracted tables, with working formulas where totals are needed.
  • A Word or PDF memo with real headings and the sources kept inside.
  • An editable PowerPoint with native charts, in one of six themes.

Working with spreadsheets and CSVs rather than prose? See AI for data analysis. Writing up research from scratch? Try the AI report generator.

Tips for better document answers

  • Ask specific questions. "What are the termination rights?" beats "summarize the contract".
  • Name the reader and length. "One page for the board" sets the level of detail.
  • Ask for quotes on key points. Rixx can quote the exact sentence alongside the page reference.
  • Split facts from interpretation. Ask Rixx to label what the document says and what it infers.
  • Use one chat per project. Follow-up questions build on the documents already loaded.

Frequently asked questions

Is there a free AI PDF summarizer in Rixx?

Yes. Every plan, including Free, gets every feature. Free includes 50 credits a day plus 500 a month. A quick cited answer uses a few credits; a deep report or a job that builds files uses more. You only pay for what a job uses. See pricing.

Can I chat with a PDF in Rixx?

Yes. Attach the PDF, ask a question and keep asking follow-ups in the same chat. Answers point back to the page they came from.

How long can a PDF be?

Up to 1,000 pages and 50 MB per file. You can upload 5 files at a time and keep up to 100 documents in one chat.

Which file types can Rixx analyse?

PDF, Word, PowerPoint, Excel, CSV, text and markdown files, and images. PDFs get page references; other documents get references back to the file.

Can Rixx compare several documents?

Yes. Upload them together or add them to the same chat, then ask for a comparison table, a memo or a workbook with one row per document. See summarize multiple PDFs.

Can Rixx pull tables out of a PDF into Excel?

Yes. Ask for an Excel workbook and Rixx extracts each table, one sheet per table, with the page it came from. In a job it checks totals with Python.

Can it turn a document into slides?

Yes. Ask for a deck and Rixx builds an editable PowerPoint with native charts and references in the speaker notes. See the AI presentation maker.