Docs
Browser & Analysis

Browser & Analysis

Live web work, research, files, and structured analysis when conversation alone is not enough.

Why This Surface Exists

Some tasks cannot be solved well from memory alone.

They need:

  • live web context
  • navigation through a real site
  • comparison across sources
  • inspection of files or spreadsheets
  • artifact generation you can actually review

That is where browser-backed work and structured analysis fit.

Browser-Backed Work

Browser-backed work matters when the truth is on a live page right now.

Examples:

  • checking a site for updated information
  • comparing product details across sources
  • gathering evidence before a recommendation
  • following a repeatable multi-step web workflow

The point is not "automation for its own sake." The point is getting to a more useful outcome when the task depends on live context.

File And Spreadsheet Analysis

A lot of real work lives inside artifacts, not chat messages.

This surface is where Viventium can help with:

  • documents
  • spreadsheets
  • exports and reports
  • structured data
  • generated summaries, tables, or charts

That matters because a second brain should be able to work with the materials your work is already made of.

Research That Feels Grounded

This layer is also where research gets better.

Instead of one confident answer from a static prompt, Viventium can:

  • compare multiple sources
  • ground conclusions in current context
  • bring findings back into the main conversation
  • attach outputs to a longer-running project when needed

What Should Stay Visible

Good browser and analysis work should stay:

  • observable
  • reviewable
  • easy to steer
  • clearly connected to the job being done

The product should never make these capabilities feel like hidden magic.

How It Works

Browser-backed work in Viventium is powered by:

  • SearXNG — self-hosted, privacy-respecting web search (no tracking, no data shared with search providers)
  • Firecrawl — deep page extraction and crawling for structured data from websites
  • Browser sandbox — within GlassHive, workers can navigate real websites through noVNC-accessible browser sessions

For quick research, the main agent or a background agent can search the web and extract information during a conversation. For deeper work that needs sustained browsing, a GlassHive worker can take over with a full browser session.

Where It Leads

Some browser and analysis work is one-off.

Some of it turns into:

  • Recurring research — scheduled web checks that run weekly or daily
  • Project artifacts — reports, comparisons, and summaries tied to a project
  • Worker responsibilities — a GlassHive worker assigned to monitor a topic or source
  • Scheduled updates — automated briefings that include fresh web context

That is why this surface connects so closely with projects and GlassHive.

Keep Reading