@capotej/pi-zsearch
v1.1.0
Published
Adds web_search and web_read tools powered by the Z.AI API.
Readme
@capotej/pi-zsearch
A pi extension that adds web search and web page reading capabilities via the Z.AI API.
Tools
web_search
Search the web using the Z.AI Web Search API. Returns structured results with titles, summaries, URLs, source names, and publication dates.
| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| query | string | ✅ | — | The search query |
| domain | string | ❌ | — | Limit to a specific domain (e.g. github.com) |
| recency | enum | ❌ | noLimit | Time filter: oneDay, oneWeek, oneMonth, oneYear, noLimit |
| content_size | enum | ❌ | medium | Summary length: medium (~400–600 words) or high (~2500 words, higher cost) |
| location | enum | ❌ | cn | Region bias: cn (Chinese region) or us (non-Chinese region) |
web_read
Read and parse a web page using the Z.AI Web Reader API. Returns page content as markdown or plain text.
| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| url | string | ✅ | — | The URL to read |
| return_format | enum | ❌ | markdown | Output format: markdown or text |
| no_cache | boolean | ❌ | false | Disable caching |
| retain_images | boolean | ❌ | true | Keep images in output |
| no_gfm | boolean | ❌ | false | Disable GitHub Flavored Markdown |
| keep_img_data_url | boolean | ❌ | false | Keep image data URLs |
| with_images_summary | boolean | ❌ | false | Include image summary |
| with_links_summary | boolean | ❌ | false | Include links summary |
| timeout | number | ❌ | 20 | Request timeout in seconds |
Installation
pi install npm:@capotej/pi-zsearchSetup
Set the ZAI_API_KEY environment variable. Get your API key at z.ai/manage-apikey.
export ZAI_API_KEY="your-api-key"Development
npm install
npm run build