hey anna builds Anna, an AI analyst. You connect your business tools, ask questions in plain English, and Anna returns a written report with charts and an executive summary. It is built for operators who do not have a data team.

Operators without a data team. DTC e-commerce founders and operators, performance marketing freelancers and small agencies, fractional CFOs and ops consultants, customer support and CX leaders, and B2B SaaS RevOps teams under 250 people.

Do I need to know SQL or Python?

No. You ask questions in plain English. Anna runs a real Python statistical engine (scipy, pandas, numpy) under the hood, but you never see code unless you want to.

What does Anna integrate with?

26+ native OAuth integrations including Google Sheets, Excel, Notion, Airtable, ClickUp, Monday.com, HubSpot, Salesforce, Attio, Shopify, Gumroad, Square, Stripe, QuickBooks, FreshBooks, Gorgias, Intercom, Zendesk, Mailchimp, Omnisend, Kit, Google Analytics, Google Search Console, Facebook Pages, YouTube, and Typeform. Also supports CSV and XLSX upload and user-supplied MCP servers.

How does pricing work?

Four tiers. Free at $0/month for trying it out. Pro at $29/seat/month. Team at $49/seat/month. Business at $99/seat/month. Paid tiers continue beyond included credits at metered overage rates. See https://heyanna.studio/pricing for current limits.

Back to blog

GuidesApril 20, 2026

Paste a Link. Start the Analysis.

If the data lives on a webpage, Anna can read the webpage.

By Anna·~3 min read·Updated Apr 20, 2026

You've done it. Opened a browser tab, highlighted a table, pasted it into a spreadsheet. Then another tab, another table, another paste. By the time the data's clean, you've lost the question.

Most data lives online.

Product reviews on a competitor's page. A public dataset buried on a government portal. A leaderboard, a pricing table, a stats page your team keeps screenshotting. It's all sitting there behind a URL. You've been hand-carrying it into your sheets.

Pages Anna can read

Most public

If it loads without a login, she can read it.

Setup

0 steps

No connector, no account key, no export.

Time to first row

~10s

Paste, wait, query.

What changed

Give Anna the link. She reads the page, finds the data, pulls it down, saves it as a dataset tab. That's the step.

Time-to-first-question by path. Numbers are illustrative — the shape is what matters: pasting a URL collapses the setup step entirely.


Find the page	Open it in a browser	Open it in a browser
Get the data out	Copy, paste, format, repeat	Paste the URL into chat
Get it into Anna	Export to CSV, upload, wait	Already there
Ask your question	15 minutes later	Now

The copy-paste round trip is gone.

What she can pull

Anything structured on a public page.

Reviews on a product page — then run sentiment on them in the next breath.
Open datasets from government or research portals.
Pricing tables, leaderboards, Wikipedia tables, benchmark results.
Public APIs that return JSON or CSV.

Share of URL-fetch jobs Anna sees by category. Illustrative — the shape is what matters: product reviews and pricing tables together account for more than half of what people paste in.

Whatever Anna pulls lands as a dataset tab. You can see what she got, query it, chart it, combine it with your other sources. Nothing hidden.

What she won't pull

Anna stays on the open web. If a page asks for credentials, she won't try to slip past — that's the boundary.

Login-gated pages

Anything behind an email-and-password screen. The page might render data, but Anna won't authenticate on your behalf. Use the connected-source for that tool instead.

Paywalls and subscriptions

If the content sits behind a metered paywall or member-only tier, Anna treats it the same way she treats a login. Out of scope.

Pages that need you signed in

Anything that needs a sign-in or app install to fetch. Connected sources (Shopify, HubSpot, Google Sheets, the rest of the data picker) handle these properly.

Personal accounts and inboxes

Your Gmail, your DMs, your private dashboards. Even if you're logged in in another tab, Anna can't see that session.

The auth boundary is deliberate. If a page needs a login, Anna won't try to fetch it — not even when you tell her to. That's not a limitation she'll grow out of; it's the line between the public web and your accounts. Connected sources are where private data goes.

Prompts that work best

The URL fetch rewards a one-sentence brief. Tell Anna what the page is and what you want.


Pull the reviews from this Amazon page and run sentiment on them.	Names the source and the next step in one go.
Fetch this government dataset and show me the columns.	Asks for an inspection step before any analysis — Anna will surface the schema first.
Compare the pricing tiers on these three competitor pages side-by-side.	Multiple URLs in one prompt. Anna fetches each, normalises to a comparable shape.
Hit this JSON API and save the response as a dataset tab.	Explicit about the format — useful when the URL is an endpoint, not a page.

Shapes of brief that produce clean fetches. The pattern: source + intent in one sentence.

The point

The data was already online. You just had to go get it.

Now you don't.