Pi Web Access

An extension for Pi coding agent that gives Pi web capabilities: search via Perplexity AI, fetch and extract content from URLs, and read PDFs.

web_search({ query: "TypeScript best practices 2025" })
fetch_content({ url: "https://docs.example.com/guide" })

Install

pi install npm:pi-web-access

Add your Perplexity API key:

# Option 1: Environment variable
export PERPLEXITY_API_KEY="pplx-..."

# Option 2: Config file
echo '{"perplexityApiKey": "pplx-..."}' > ~/.pi/web-search.json

Get a key at https://perplexity.ai/settings/api

Requires: Pi v0.37.3+

Tools

web_search

Search the web via Perplexity AI. Returns synthesized answer with source citations.

// Single query
web_search({ query: "rust async programming" })

// Multiple queries (parallel)
web_search({ queries: ["query 1", "query 2"] })

// With options
web_search({
  query: "latest news",
  numResults: 10,              // Default: 5, max: 20
  recencyFilter: "week",       // day, week, month, year
  domainFilter: ["github.com"] // Prefix with - to exclude
})

// Fetch full page content (async)
web_search({ query: "...", includeContent: true })

When includeContent: true, sources are fetched in the background. Agent receives notification when ready.

fetch_content

Fetch URL(s) and extract readable content as markdown.

// Single URL - returns content directly (also stored for retrieval)
fetch_content({ url: "https://example.com/article" })

// Multiple URLs - returns summary (content stored for retrieval)
fetch_content({ urls: ["url1", "url2", "url3"] })

// PDFs - extracted and saved to ~/Downloads/
fetch_content({ url: "https://arxiv.org/pdf/1706.03762" })
// → "PDF extracted and saved to: ~/Downloads/arxiv-170603762.md"

PDF handling: When fetching a PDF URL, the extension extracts text and saves it as a markdown file in ~/Downloads/. The agent can then use read to access specific sections without loading 200K+ chars into context.

get_search_content

Retrieve stored content from previous searches or fetches.

// By response ID (from web_search or fetch_content)
get_search_content({ responseId: "abc123", urlIndex: 0 })

// By URL
get_search_content({ responseId: "abc123", url: "https://..." })

// By query (for search results)
get_search_content({ responseId: "abc123", query: "original query" })

Features

Activity Monitor (Ctrl+Shift+O)

Toggle live request/response activity:

─── Web Search Activity ────────────────────────────────────
  API  "typescript best practices"     200    2.1s ✓
  GET  docs.example.com/article        200    0.8s ✓
  GET  blog.example.com/post           404    0.3s ✗
  GET  news.example.com/latest         ...    1.2s ⋯
────────────────────────────────────────────────────────────
Rate: 3/10 (resets in 42s)

RSC Content Extraction

Next.js App Router pages embed content as RSC (React Server Components) flight data in script tags. When Readability fails, the extension parses these JSON payloads directly, reconstructing markdown with headings, tables, code blocks, and links.

TUI Rendering

Tool calls render with real-time progress:

┌─ search "TypeScript best practices 2025" ─────────────────────────┐
│ [████████░░] searching                                            │
└───────────────────────────────────────────────────────────────────┘

Commands

/search

Browse stored search results interactively.

How It Works

Agent Request → Perplexity API → Synthesized Answer + Citations
                                         ↓
                              [if includeContent: true]
                                         ↓
                              Background Fetch (3 concurrent)
                                         ↓
                        ┌────────────────┼────────────────┐
                        ↓                ↓                ↓
                       PDF          HTML/Text          RSC
                        ↓                ↓                ↓
                   unpdf →        Readability →    RSC Parser →
                 Save to file      Markdown          Markdown
                        ↓                ↓                ↓
                        └────────────────┼────────────────┘
                                         ↓
                              Agent Notification (triggerTurn)

Rate Limits

• Perplexity API: 10 requests/minute (enforced client-side)
• Content Fetch: 3 concurrent requests, 30s timeout per URL
• Cache TTL: 1 hour

Files

| File | Purpose | |------|---------| | index.ts | Extension entry, tool definitions, commands, widget | | perplexity.ts | Perplexity API client, rate limiting | | extract.ts | URL fetching, content extraction routing | | pdf-extract.ts | PDF text extraction, saves to markdown | | rsc-extract.ts | RSC flight data parser for Next.js pages | | storage.ts | Session-aware result storage | | activity.ts | Activity tracking for observability widget |

Limitations

• Content extraction works best on article-style pages
• Heavy JS sites may not extract well (no browser rendering), though Next.js App Router pages with RSC flight data are supported
• PDFs are extracted as text (no OCR for scanned documents)
• Max response size: 20MB for PDFs, 5MB for HTML
• Max inline content: 30,000 chars per URL (larger content stored for retrieval via get_search_content)
• Requires Pi restart after config file changes