Browser
Your agent can visit live websites in its own cloud browser — read dashboards, extract data, fill multi-step flows, and capture screenshots and PDFs.
A lot of the information that matters in crypto never shows up in an API: a protocol's live dashboard, a governance vote, a docs page that changed yesterday. The Browser capability lets your agent open those pages itself, in its own cloud browser, and work with what it finds.
You don't operate the browser — you just ask. "Check the current parameters on this governance proposal." "Pull the numbers from this dashboard." Your agent opens the page, reads it, and brings the answer back into the conversation.
What it can do
- Read JS-heavy pages. Modern dashboards and governance UIs render in the browser, not in raw HTML. Your agent sees the page the way you would.
- Extract data. Tables, metrics, vote tallies, parameter values — pulled off the page and into its answer.
- Work through multi-step flows. It can click through a site — open a page, apply a filter, drill into a detail view — when the information takes more than one step to reach.
- Capture screenshots. Visual proof of what a page showed at the time it looked.
- Save pages as PDFs. Useful when you want a whole page preserved, not just the numbers.
What you see in chat
Browser work is visible, not a black box. While your agent browses, the chat shows a live session timeline with screenshots, so you can watch what it's looking at as it works.
When the session ends, it publishes artifacts you can open directly from the conversation:
| Artifact | What it is |
|---|---|
| Screenshot | An image of the page at a moment in the session |
| A saved copy of a full page | |
| Data | Extracted values from the page, in a form you can reuse |
When your agent reaches for it
You rarely need to say "use the browser." Your agent picks it up when a question needs a live page — typical cases:
- A protocol dashboard whose numbers you want checked right now
- A governance proposal you want summarized before the vote closes
- A project site or docs page you want compared against what was there before
- Anything a normal web search returns as a link rather than an answer
If a plain web search or a data source can answer faster, it uses that instead — the browser is for pages that have to be visited.
Usage
Browser sessions are heavier than ordinary chat: keeping a live browser open and stepping through pages draws more of your monthly usage than a quick question does. That's usually fine for a handful of sessions, but worth knowing if you ask for a lot of deep page work. See Usage for how usage works and where to check what you have left.
Browser is a built-in capability, managed like any other from the My Agent page at https://blockmind.app/elite/agent.
Whatever your agent finds in the browser, it's research material — context for your decisions, not financial advice. Your agent will never tell you what to buy or sell.
Related
KOL analysis
Get market analysis in a specific analyst's voice and framework — single-expert Q&A, two-framework synthesis, or a full expert discussion.
Charts & dashboard
Ask for interactive price charts right in chat, and shape your Elite home with up to 11 widgets — by hand or just by asking your agent.