Your web page
Paste your web page. Nothing to upload.
Make a financial analysis
The narrative section and the table on the same page frequently describe different quarters.
Free to try, no card. Your web page opens in Ancher.
Paste your web page. Nothing to upload.
Every one of these stays separately addressable instead of collapsing into one block of text.
Figures checked against the sentence or table that supports them.
Results published as web pages lead with a paragraph of adjectives and put the numbers in a table further down, and the two do not always describe the same thing on the same basis. Ancher keeps both, reads the tables including the ones published as images, and reports where the narrative's characterisation and the figures diverge instead of choosing between them.
The honest version
Strong performance in the highlights can sit above a table showing a decline on a like-for-like basis, with the difference accounted for by an acquisition mentioned in a footnote. Both are on the same page and only one survives a text extraction that flattens tables. Keeping the table and the images is what makes the comparison possible at all.
One screen of one page, at table 2, read two ways.Illustrative. The figure is invented; the gap it falls into is not.
A strong quarter with revenue growth of across the group
Table further down the page — not readThe specific never reaches the draft
Lands in the draft, with its timecode
What comes out
Figures checked against the sentence or table that supports them.
What the figures amount to, stated before the company's own framing.
Each figure with its table, its basis and its URL.
Period on period, flagging where the basis changed.
Caveats, adjustments and anything the page discloses in a footnote.
The page, its date, and whether a filing exists to check it against.
The link on its own gets a summary. This is the instruction that produces the 5 sections above, in that order, with the rules that keep them honest. Paste it with your web page — in Ancher, or in whatever assistant you already use.
Analyse the figures on this results page. State what they amount to before repeating the company's framing. List every figure with the table it came from and the basis it is stated on — reported, adjusted, organic, pro-forma — including figures inside table images. Flag where the narrative characterisation and the tables diverge. Note any change of basis between periods, and link the filing where the page provides one.
The detail, if you want it
| From the web page | Into | Why |
|---|---|---|
| Inline images | Evidence per figure | Results tables are routinely published as images, and everything that qualifies a headline number lives inside them. |
| Readable content, decluttered | Headline figures | Stripping the page down separates the company's narrative framing from the figures it is framing. |
| Title, author, published date | Sources | A results page's date determines which period it covers, and investor sites keep old releases live indefinitely. |
| Outbound links | Sources | Investor pages link to the filing, which is the document that should settle anything material. |
| Many pages in one collection | Trend | Several quarters of releases in one collection is what makes a change of basis visible, since each page alone looks consistent. |
Navigation, ads and cookie banners are dropped, so the collection holds the article rather than the page.
A date is what stops a three-year-old post from being read as current.
What a writer links to repeatedly reveals the sources they actually trust.
Charts and diagrams in an article are usually the evidence the prose is describing.
A single page is an anecdote; the value of this route appears once several are compared.
Yes, inline images are captured. That matters here more than almost anywhere else, because results tables are frequently images and they hold the basis on which every headline figure was calculated.
Both are reported with the divergence stated. Choosing one silently is how an analysis ends up repeating a company's framing while appearing to have checked it.
For a fast read, yes, with everything traceable to the URL. For a decision, follow the link to the filing — the analysis flags where one exists and what should be verified against it.
Everything you save lives in one workspace, so the financial analysis is built from your sources — not from a model's memory of the internet.
Open Ancher →Not a web page? The same financial analysis also comes from YouTube video, X thread, PDF.