Add Evelant Extension
Add to ChromeRead the whole page
The full document, including what sits below the fold and inside long PDFs.
Ask the page you are already on.
Evelant reads the whole document, including the part below the fold, and keeps a separate conversation in every tab you open.
Ask the page you are already on
- The full document, including what sits below the fold
- Long PDFs read in place, without uploading them anywhere
- No copying, no switching tab, no pasting it into a chat box
Page context
Works in Chrome and every Chromium browser
Chrome
Edge
Brave
Helium
Opera
VivaldiDifferent work, same shortcut.
Everything happens on the page you already opened.
No copying into another tab, no uploading a file you already have open.
Keep a thread per tab
Three sources open means three conversations, none overwriting another.
Check what is recent
Answers built from pages fetched at that moment, each one listed underneath.
Switch model per message
Gemini, ChatGPT and Claude, chosen at the moment you send.
The parts that are not standard.
Reading something with another person, real files out of a sandbox, a question you can say out loud, and the option of your own models.
2 reading this
What does this page actually claim?
Answering both
On Pro
Read it with someone else
Invite someone into the thread and the same answer streams to both of you at once.
Writes the PythonRuns it in a sandboxreport.pdf
Ask for the file, and get the file
It writes the code, runs it in a sandbox, and hands back a real PDF, spreadsheet or deck — not a description of one.
Alt+S
Ask without typing
Hold Alt+S, say it out loud, and the answer comes back on the orb.
DeepSeekKimiQwenOllama
No credits spentOn Max
Or run it on your own key
DeepSeek, Kimi, Qwen or a local Ollama model, billed to you and costing no credits here.
See it on a page before you install it.
Both shapes the extension takes, with every control it actually ships.
Summarise this page and flag what it does not cover.
The page explains how the process works and what it produced, but never states how the sample was chosen or how large it was.
Answered from the page you are onSummarise this page and flag what it does not cover.
Appearance
Theme
Position
Ctrl+B
Open the chat
Brings the chatbox up on the page you are reading, and the same keys put it away.
Reads the page you are on
The answer comes from the article in front of you, and says which part of it the answer came from.
Carries your open tabs
Scan a few pages and ask across all of them at once, without pasting anything between them.
Three more things it does without leaving the page.
Selections, attachments, and the sources behind an answer.
The orb carries the four things you would otherwise open a chat to ask for: capture a region, summarise the page, ask about the text you highlighted, or type a quick question.
Each one arrives with the page already attached, so the question is the only thing you have to write.
The orb
On any page
Capture a region
Summarise this page
Ask the selection
Sync everywhere.
Ask on the page you are reading, and pick the same thread up on your phone.
On your phone
The thread goes with you
The web app opens on the same conversation the extension was in, already scrolled to where you left it. Nothing to copy across, nothing to re-ask.
Mobile
On the page
Open where you are reading
The panel sits beside the article rather than on top of it, so the page you are asking about stays in front of you while the answer arrives.
Page
Panel
How an answer gets built.
Read where you are, split into passages, ranked against what you asked.
- This page
Read where you are
- Evelant
Finds what matters, using:
- Passages · Page split and indexed
- Ranking · Matched to your question
- Model · Gemini, ChatGPT, Claude
- Web search · When the page falls short
- The answer
With its source named
Read once, kept as passages
The page is split and indexed, so a question can be matched against the part that answers it rather than against the whole document.
A thread for every tab
Each tab carries its own conversation and its own index of the page in it. Nothing bleeds from one into another.
Three providers, in order
A model that is slow or down hands the message to the next one in the chain, which is why a bad afternoon at one provider is not an outage here.
Free to begin, with no card.
Start without an account, and pay only once the daily limit is the thing slowing you down.
Free
$0
foreverEverything you need to get started.
- Fifteen messages a day, free forever
- Fast models from Gemini, ChatGPT and Claude
Create an account
No card, and no account needed to try it.
ProMost popular
$11.24
/ mo$14.99You save $45 a year, 25% off
For everyday AI users who want every model.
- Everything in Free, without the daily cap
- A thousand credits a month — one for a fast message, up to six for the largest
- Every model, chosen per message
- Deep context on long documents, at a higher credit rate
- Shared rooms, for reading something with someone else
Upgrade to Pro
$134.88 billed once a year. Cancel anytime.
MaxPower
$69
/ moFor heavy users — the largest allowance, and your own key if you want it.
- Everything in Pro
- 5,000 credits per month
- Every model, with the largest credit allowance
- Bring your own API key (BYOK)
- Priority traffic
Get Max
$69 billed monthly. Cancel anytime.
Frequently asked questions
Still wondering about something? We would love to hear from you.
Click the question to see the answer.
No. It reads a page when you open the assistant on it, and not before. There is no background pass over your browsing, and nothing leaves the tab until you send a question.
To whichever model answers it — Gemini, ChatGPT or Claude — and to your own history, so the thread is there when you come back. You can clear one session or all of them from Settings.
One message on a fast model. The heavier models cost more per message, up to six for the largest, and reading a full long document costs more again. The model picker shows the price before you send.
Ten questions a day with no account, fifteen with one, on the fast Gemini, ChatGPT and Claude models. No card at any point.
Yes. A thousand credits a month, every model including the reasoning ones, and no daily cap. Cancelling from the billing portal keeps Pro live until the period you have already paid for runs out.
Yes, on a paid plan. Invite them into a session and you both watch the same answer stream in, whether you are on the page or in the web app.
Yes. It reads what your browser is already showing, so an internal wiki, a dashboard or a subscription article is no different to it than a public page.
No. Keys are optional and only on Max, where the messages that run on your key cost no credits at all.
Get started.
A minute to install, and no card at any point.
Sign in to keep it
Ten a day without an account, fifteen and a kept history with one.
Then it is just there
The orb follows your cursor, and these two keys open the chat.
Ctrl+B