> ## Documentation Index
> Fetch the complete documentation index at: https://hyperfx.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Browser Contexts

> Save authenticated browser sessions so agents can take actions on the web on your behalf — log into LinkedIn, your ad platforms, dashboards, anywhere.

A **browser context** is a saved, authenticated browser session that an agent can use when it needs to take actions on the web. You log in once, save the session, and agents can use it for automated browsing without re-authenticating each time.

## When to use this

Use a browser context when:

* The site has no API or your agent doesn't have an integration for it
* You want the agent to act *as you* on a site (your LinkedIn profile, your ad account, your CRM, your CMS)
* The work involves authenticated dashboards, reports, or platforms that aren't covered by Hyper's [native integrations](/integrations/overview)

If a native integration exists, prefer that — they're faster and more reliable. Browser contexts are the catch-all when there isn't one.

<Info>
  **Web Browsing vs Web Search.** Browser contexts power the Web Browsing tool — agents interacting with logged-in pages and maintaining sessions. This is different from Web Search, which queries the public web without authentication. Web Search doesn't need a browser context.
</Info>

## Creating a browser context

<Steps>
  <Step title="Open Browser Contexts">
    From the left sidebar, click **More** → **Browser Contexts**.
  </Step>

  <Step title="Create new">
    Click **Create new context**, name it (e.g. "LinkedIn — recruiting", "Meta Business Manager — main account"), and select the site.
  </Step>

  <Step title="Log in">
    Hyper opens the site in a sandboxed browser. Log in like you normally would — username, password, MFA, whatever's needed.
  </Step>

  <Step title="Save the session">
    Once you're logged in, save. The authenticated session is stored encrypted and ready for agents to use.
  </Step>
</Steps>

## Using a browser context in chat

Just tell the agent which context to use:

```text theme={null}
Use my LinkedIn browser context to find recent posts from people working
on AI agents at YC startups, and DM the top 5 most engaged ones.
```

Agents can also pick a context automatically based on the site and the task — but being explicit is faster and more reliable.

You can also pin a browser context to an agent's **Resources** tab, so it's the default for that agent.

## Common use cases

* **LinkedIn** — outbound, profile updates, content posting, connection management
* **Ad platforms** — scraping reports out of Meta Business Manager, Google Ads, TikTok Ads when an integration isn't enough
* **Internal dashboards** — pulling metrics out of Mixpanel, Looker, internal admin tools
* **CMSs and ops tools** — Webflow, Notion, Airtable, project trackers
* **Anywhere a login is required** and there's no native integration

## Security

* **Encrypted at rest** — sessions are stored encrypted, no plain-text credentials
* **Private to you** — contexts are tied to your user, not shared across the workspace by default
* **Revokable** — delete a context any time; sessions also expire on their own for safety
* **Sandboxed browser** — agent browsing happens in an isolated environment, separate from your machine

If you want to share a context with teammates, use a shared service account login when you create the context.

## Going further

* Combine a browser context with a [scheduled task](/agents#tasks) to run authenticated workflows on a cadence
* Pin contexts to a specific agent in its [Resources tab](/agents#the-resources-tab)
* For sites with native support, browse [Integrations](/integrations/overview) before reaching for a browser context