Getting started

If you just want to try the package in Pi, this is the short version.

Install it

pi install npm:@demigodmode/pi-web-agent

If Pi is already running, reload or restart it so the package gets picked up.

What you get

The package exposes four tools:

• web_search
• web_fetch
• web_fetch_headless
• web_explore

The important part is that they do different jobs on purpose.

web_search finds sources.

web_fetch tries to read one page over plain HTTP.

web_fetch_headless is there for pages that really do need a browser.

web_explore is the higher-level research path when the job is bigger than one search or one page read.

Try a few prompts

Search for sources

Search for recent docs about Playwright browser installation.

That should push Pi toward web_search when the job is discovery.

Read one page directly

Fetch and summarize https://example.com/article

That should favor web_fetch when the job is reading a specific page over HTTP.

Use the browser path explicitly

Use headless fetch to read https://example.com/app

That should favor web_fetch_headless when browser rendering is actually the point.

Let the package do bounded research

Find current docs and discussions about configuring Vitest coverage in TypeScript projects.

That should favor web_explore for the broader research workflow.

Presentation defaults

The package renders web tool output in compact mode by default.

If you want to change that, open the settings UI with:

/web-agent

That lets you change the default mode and per-tool overrides without editing code.

If you want the full details, see Presentation and settings.

One thing to keep in mind

If the package says a page was too weak to trust over HTTP, that is not it being stubborn. That is the whole point of the contract.