@xaccefy/pi-lookup
Web search, page fetch, library docs (Context7), and GitHub repo Q&A (DeepWiki) for Pi Agent
Package details
Install @xaccefy/pi-lookup from npm and Pi will load the resources declared by the package manifest.
$ pi install npm:@xaccefy/pi-lookup- Package
@xaccefy/pi-lookup- Version
0.5.4- Published
- Jul 18, 2026
- Downloads
- 3,441/mo · 656/wk
- Author
- xaccefy
- License
- MIT
- Types
- extension
- Size
- 41.1 KB
- Dependencies
- 2 dependencies · 3 peers
Pi manifest JSON
{
"extensions": [
"./src/index.ts"
]
}Security note
Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.
README
pi-lookup
Web search, page fetch, library docs (Context7), and GitHub Q&A (DeepWiki) for Pi Agent. No API keys for search/fetch/docs (DeepWiki is public repos only).
Install
pi install npm:@xaccefy/pi-lookup
Tools
web_search
Asks search engines through the local open-websearch tool (no API key).
query(required)limit(optional, default 10)engines(optional, e.g.duckduckgo,brave,bing)
web_fetch
Fetches a URL as clean text/markdown (GitHub READMEs use a dedicated path).
url(required, http/https)
JS pages: if it only sends back a bare HTML shell, web_fetch re-draws it with system Chromium (--headless --dump-dom) when present, and uses the fuller text.
| Variable | Purpose |
|---|---|
PI_CHROMIUM_PATH |
Absolute path to chromium/chrome binary |
PI_WEBSEARCH_PORT |
Daemon port (default 3210) |
It also tries: /usr/bin/chromium, /usr/sbin/chromium, Chrome stable paths, etc. If Chromium isn’t there, you get the plain extract.
context7
Up-to-date library docs + examples.
libraryName(required, e.g.react)topic(optional, e.g.hooks)maxTokens(optional, default 10000)
deepwiki
Natural-language Q&A over a public GitHub repo.
repo(required,owner/name)question(required)
Lifecycle
- session_start: tries to start the
open-websearchtool (non-blocking). - session_shutdown: stops the tool this add-on started.
Development
bun test packages/pi-lookup
bun run typecheck