@apmantza/greedysearch-pi 1.1.2 → 1.1.4

package/README.md

@@ -1,70 +1,169 @@
-# GreedySearch for Pi
-
-Pi extension that adds a `greedy_search` tool — fans out queries to Perplexity, Bing Copilot, and Google AI simultaneously and returns synthesized AI answers. Gemini can act as an optional synthesizer via `synthesize: true`, deduplicating sources across engines and returning a single grounded answer.
-
-Forked from [GreedySearch-claude](https://github.com/apmantza/GreedySearch-claude).
-
-## Install
-
-```bash
-pi install npm:@apmantza/greedysearch-pi
-```
-
-Or directly from git:
-
-```bash
-pi install git:github.com/apmantza/GreedySearch-pi
-```
-
-## Usage
-
-Once installed, Pi gains a `greedy_search` tool. The model will use it automatically for questions about current libraries, error messages, version-specific docs, etc.
-
-You can also invoke it directly:
-
-```
-greedy_search({ query: "best way to handle auth in Next.js 15", engine: "all" })
-```
-
-With Gemini synthesis:
-
-```
-greedy_search({ query: "best way to handle auth in Next.js 15", engine: "all", synthesize: true })
-```
-
-**Engines:**
-- `all` — fan-out to all three in parallel (default, highest confidence)
-- `perplexity` / `p` — best for technical Q&A
-- `bing` / `b` — best for recent news and Microsoft ecosystem
-- `google` / `g` — best for broad coverage
-- `gemini` / `gem` — Gemini standalone query
-
-**`synthesize: true`**
-
-Deduplicates sources across engines by consensus, feeds them to Gemini, and returns a single grounded answer instead of three separate responses. Adds ~30s to the request but reduces downstream token usage when passing results to a model.
-
-## Requirements
-
-- Chrome must be running (or it auto-launches a dedicated instance via `launch.mjs`)
-- The `chrome-cdp` skill must be accessible (same CDP infrastructure as GreedySearch-claude)
-
-## Setup (first time)
-
-To pre-launch the dedicated GreedySearch Chrome instance:
-
-```bash
-node ~/.pi/agent/git/GreedySearch-pi/launch.mjs
-```
-
-Stop it when done:
-
-```bash
-node ~/.pi/agent/git/GreedySearch-pi/launch.mjs --kill
-```
-
-## How It Works
-
-- `index.ts` — Pi extension, registers `greedy_search` tool
-- `search.mjs` — CLI runner, spawns extractors in parallel
-- `launch.mjs` — launches dedicated Chrome on port 9223
-- `extractors/` — per-engine CDP scrapers (Perplexity, Bing Copilot, Google AI, Gemini)
+# GreedySearch for Pi
+
+Pi extension that adds a `greedy_search` tool — fans out queries to Perplexity, Bing Copilot, and Google AI simultaneously and returns AI-synthesized answers with deduped sources. Streams progress as each engine completes.
+
+Forked from [GreedySearch-claude](https://github.com/apmantza/GreedySearch-claude).
+
+## Install
+
+```bash
+pi install npm:@apmantza/greedysearch-pi
+```
+
+Or directly from git:
+
+```bash
+pi install git:github.com/apmantza/GreedySearch-pi
+```
+
+## Quick Start
+
+Once installed, Pi gains a `greedy_search` tool. The model will use it automatically for questions about current libraries, error messages, version-specific docs, etc.
+
+```
+greedy_search({ query: "What's new in React 19?", engine: "all" })
+```
+
+## Parameters
+
+| Parameter | Type | Default | Description |
+|-----------|------|---------|-------------|
+| `query` | string | required | The search question |
+| `engine` | string | `"all"` | Engine to use (see below) |
+| `synthesize` | boolean | `false` | Synthesize results into one answer via Gemini |
+| `fullAnswer` | boolean | `false` | Return complete answer (~3000+ chars) vs truncated preview (~300 chars) |
+
+## Engines
+
+| Engine | Alias | Latency | Best for |
+|--------|-------|---------|----------|
+| `all` | — | 30-90s | Highest confidence — all 3 engines in parallel (default) |
+| `perplexity` | `p` | 15-30s | Technical Q&A, code explanations, documentation |
+| `bing` | `b` | 15-30s | Recent news, Microsoft ecosystem |
+| `google` | `g` | 15-30s | Broad coverage, multiple perspectives |
+| `gemini` | `gem` | 15-30s | Google's AI with different training data |
+
+## Streaming Progress
+
+When using `engine: "all"`, the tool streams progress as each engine completes:
+
+```
+**Searching...** ⏳ perplexity · ⏳ bing · ⏳ google
+**Searching...** ✅ perplexity done · ⏳ bing · ⏳ google
+**Searching...** ✅ perplexity done · ✅ bing done · ⏳ google
+**Searching...** ✅ perplexity done · ✅ bing done · ✅ google done
+```
+
+## Synthesis Mode
+
+For complex research questions, use `synthesize: true` with `engine: "all"`:
+
+```
+greedy_search({ query: "best auth patterns for SaaS in 2026", engine: "all", synthesize: true })
+```
+
+This deduplicates sources across engines and feeds them to Gemini for one clean, synthesized answer. Adds ~30s but produces the highest quality output with deduped sources showing consensus scores (`[2/3]`, `[3/3]`).
+
+**Use synthesis when:**
+- You need one definitive answer, not multiple perspectives
+- You're researching a topic to write about or make a decision
+- Token efficiency matters (one answer vs three)
+
+**Skip synthesis when:**
+- You want to see where engines disagree
+- Speed matters
+
+## Full vs Short Answers
+
+Default mode returns ~300 char summaries to save tokens. Use `fullAnswer: true` for complete responses:
+
+```
+greedy_search({ query: "explain the React compiler", engine: "perplexity", fullAnswer: true })
+```
+
+## Examples
+
+**Quick technical lookup:**
+```
+greedy_search({ query: "How to use async await in Python", engine: "perplexity" })
+```
+
+**Compare tools (see where engines agree/disagree):**
+```
+greedy_search({ query: "Prisma vs Drizzle in 2026", engine: "all" })
+```
+
+**Research with synthesis:**
+```
+greedy_search({ query: "Best practices for monorepo structure", engine: "all", synthesize: true })
+```
+
+**Debug an error:**
+```
+greedy_search({ query: "Error: Cannot find module 'react-dom/client' Next.js 15", engine: "all" })
+```
+
+## Requirements
+
+- **Chrome** — must be installed. The extension auto-launches a dedicated Chrome instance on port 9222 (separate from your main browser session).
+- **Node.js 22+** — for built-in `fetch` and WebSocket support.
+
+## Setup (first time)
+
+To pre-launch the dedicated GreedySearch Chrome instance:
+
+```bash
+node ~/.pi/agent/git/GreedySearch-pi/launch.mjs
+```
+
+Stop it when done:
+
+```bash
+node ~/.pi/agent/git/GreedySearch-pi/launch.mjs --kill
+```
+
+Check status:
+
+```bash
+node ~/.pi/agent/git/GreedySearch-pi/launch.mjs --status
+```
+
+## Troubleshooting
+
+### "Chrome not found"
+Set the path explicitly:
+```bash
+export CHROME_PATH="/path/to/chrome"
+```
+
+### "CDP timeout" or "Chrome may have crashed"
+Restart GreedySearch Chrome:
+```bash
+node ~/.pi/agent/git/GreedySearch-pi/launch.mjs --kill
+node ~/.pi/agent/git/GreedySearch-pi/launch.mjs
+```
+
+### Google / Bing "verify you're human"
+The extension auto-clicks simple verification buttons. For CAPTCHAs, solve manually in the Chrome window that opens.
+
+### Search hangs
+Chrome may be unresponsive. Restart it with `launch.mjs --kill` then `launch.mjs`.
+
+### Sources are junk links
+This was a known issue with Gemini sources. If you're on an older version, update:
+```bash
+pi install npm:@apmantza/greedysearch-pi
+```
+
+## How It Works
+
+- `index.ts` — Pi extension, registers `greedy_search` tool with streaming progress
+- `search.mjs` — CLI runner, spawns extractors in parallel, emits `PROGRESS:` events to stderr
+- `launch.mjs` — launches dedicated Chrome on port 9222 with isolated profile
+- `extractors/` — per-engine CDP scrapers (Perplexity, Bing Copilot, Google AI, Gemini)
+- `cdp.mjs` — Chrome DevTools Protocol CLI for browser automation
+- `skills/greedy-search/SKILL.md` — skill file that guides the model on when/how to use greedy_search
+
+## License
+
+MIT

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@apmantza/greedysearch-pi",
-  "version": "1.1.2",
+  "version": "1.1.4",
   "description": "Pi extension: search Perplexity, Bing Copilot, and Google AI in parallel with optional Gemini synthesis — grounded AI answers, not just links",
   "type": "module",
   "keywords": [

package/results/2026-03-21_11-35-35_test-query-for-file-writing.json

@@ -1,6 +0,0 @@
-{
-  "query": "test query for file writing",
-  "url": "https://www.perplexity.ai/search/test-query-for-file-writing-.FobTp2uSWSg70txH6RthQ",
-  "answer": "This message is just a test query and does not trigger any actual file‑writing action on your system.  \n\nIf you want, you can describe what kind of file you’d like to write (for example: a text file, Python script, CSV, etc.), and I can give you a concrete example in code that you can run…",
-  "sources": []
-}