How do I know the numbers it gives me are right?

Every figure in the brief links back to the exact filing and page the agent read it from. Click through and check it against the source yourself. The agent opens the actual 10-K, 10-Q, or transcript and reads the rendered tables, so the number comes from the document rather than a paraphrase. If a figure can’t be sourced, it’s flagged rather than guessed.

Can it use my paid research access, like a brokerage portal or terminal?

Yes. The agent drives the Chrome you’re already signed into, so it reads the brokerage research portals, data terminals, and IR viewers you have a seat for, all in your own session. Your credentials and cookies stay in your browser; Actionbook never stores or re-indexes the content. You operate within your own license, so honor your provider’s automation terms. The agent does what you could do by hand.

Is this financial advice?

No. Actionbook retrieves and organizes primary sources. It opens the filings, extracts the figures, and shows you where each one came from. It doesn’t recommend trades or take positions. The reading, the judgment, and every decision stay yours; the agent just gets the documents and the numbers in front of you faster.

How fresh is the data it pulls?

It reads live EDGAR full-text search and investor-relations pages at the moment you run the prompt, so you’re never stuck at a model’s training cutoff. If a company filed an 8-K this morning, the agent opens this morning’s filing. That’s the point of opening the real document instead of recalling a cached one.

Which AI agents can do financial report search through Actionbook?

Actionbook works with ChatGPT, Claude, Gemini, Perplexity, OpenClaw, Hermes, and any MCP-compatible client. Connect the one you already use and it can run EDGAR searches, open the actual filings, and extract the figures in your own Chrome, each one sourced to its page.

Financial report search

Opens the filing, not just a snippet

Your AI pulls 10-Ks, 10-Qs, and transcripts from EDGAR, IR pages, and your research portals, citing the exact page behind each number.

Start free See example prompts

sec.gov/edgar/browse?CIK=NTRN

Pulling the 10-K from EDGARChatGPT, Claude or Gemini

“NTRN: pull segment revenue and YoY from the latest 10-K, with a page reference for every figure.”

SegmentRevenueYoY

Cloud

10-K p.47

$4.2B+18%

Hardware

10-K p.49

$2.8B+6%

Services

10-K p.52

$1.1B-3%

Reading the filing…Every figure sourced

[ 01 ]How it works

From a ticker to a sourced brief

Connect the AI you already use, whether that’s ChatGPT, Claude, or Gemini, to your browser. From a single prompt it runs four steps, and every figure ends up sourced to its page.

1
Point your AI at the primary sources
Link ChatGPT, Claude, or Gemini over MCP in a couple of clicks. It gains a real browser: the same Chrome where you already read EDGAR, IR pages, and your research portals, logins included.
2
Name the company and the question
Give it a ticker and what you need: segment revenue, a quarter-over-quarter diff, a risk-factor change. It opens the actual filing instead of paraphrasing a cached snippet.
3
It reads the document, not a summary of one
It opens the 10-K, 10-Q, 8-K, or transcript and reads the rendered tables straight from the source: segment breakdowns, cash flow, footnotes, restatements.
Summary-only chatbots never open the filing they’re quoting.
4
It returns a brief where every number is sourced
You get the figures organized the way you asked, and each one carries the exact document and page it came from. Click through and verify it yourself.

[ 02 ]Every figure cited

docs.google.com/spreadsheets

Awaiting approval

Your AI wants to export the figures table to your model sheet.

model-sheet · NTRN segments

SegmentRevenueYoY

Cloud

10-K p.47

$4.2B+18%

Hardware

10-K p.49

$2.8B+6%

Services

10-K p.52

$1.1B-3%

Export the figures. You approve first.

When you ask the agent to write extracted figures into your model sheet, that’s a real action. So it works like every consequential Actionbook step: you see exactly what it wants to do before it commits, in plain language.

See it before it writes: Exports and downloads surface for your approval first, with the target spelled out.
Sourced on the way out: The figures carry their filing-and-page citations into the sheet, not just bare numbers.
Watch it work: Everything happens in your own Chrome tab, so you can take over at any point.

Start free

[ 03 ]Who it's for

Built for how people read filings

One workflow, whether you’re screening, tracking, or fact-checking.

Equity analysts

Run a first-pass screen straight from the latest 10-K, covering segment revenue, margins, and guidance, with every figure traceable to the page it sits on.

Page-cited figures

Cloud10-K p.47$4.2B+18%

Hardware10-K p.49$2.8B+6%

Services10-K p.52$1.1B-3%

Corporate finance teams

Track competitors’ filings as they post: pull the new 8-K, diff the last two 10-Qs, and build a clean comparison without copy-pasting tables by hand.

Retail investors

Do real homework before you buy. Read what the company actually filed instead of a chatbot’s paraphrase, with the source page one click away.

The real document, not a summary

It opens the actual 10-K, 10-Q, or transcript in your own browser and reads the rendered tables. You get the primary source, not summaries of summaries of an indexed extract.

Journalists

Verify a claim against the primary document. The agent opens the filing or transcript and shows you the exact passage the number came from.

[ 04 ]Copy a prompt

Prompts to start researching

Copy a prompt, paste it into your AI, and Actionbook runs it in your own browser session.

Segment revenue from the 10-K

Open the latest 10-K for NTRN on SEC EDGAR, pull segment revenue and year-over-year growth, and return a table where every figure cites the exact filing page it came from.

Diff the last two 10-Qs

Find the two most recent 10-Q filings for this company on EDGAR and tell me what changed between them, focusing on risk factors and forward guidance, with a page reference for each change.

3-company comparison table

Build a comparison table of three companies from their latest filings: revenue, gross margin, and operating income. Open each filing on EDGAR and link the source page next to every number.

Pricing in the earnings call

Open the latest earnings-call transcript on the investor-relations page and find every mention of pricing or price increases. Quote each passage and link where it appears.

Check the latest 8-K

Pull the most recent 8-K this company filed on SEC EDGAR, tell me whether anything in it is material, and for each item give the item number, a one-line summary, and the filing date.

From my research portal

Open my logged-in brokerage research portal, find the latest analyst note on this ticker, and summarize the rating, price target, and key thesis, with a link back to the note. Ask before downloading anything.

Frequently asked questions about financial report search

What analysts ask before letting an AI pull figures from their filings and portals.

Yes. The agent drives the Chrome you’re already signed into, so it reads the brokerage research portals, data terminals, and IR viewers you have a seat for, all in your own session. Your credentials and cookies stay in your browser; Actionbook never stores or re-indexes the content. You operate within your own license, so honor your provider’s automation terms. The agent does what you could do by hand.