openbuilder 0.1.0 → 0.1.2

package/README.md

@@ -293,14 +293,78 @@ To set up OpenBuilder as an automated meeting bot (e.g. for OpenClaw agents):
 - Caption mode (`--captions`) is the most reliable on headless servers
 - Audio mode (`--audio`) requires PulseAudio + ffmpeg + OpenAI key + Xvfb (experimental on servers)
 
+## Platform Compatibility
+
+OpenBuilder works anywhere OpenClaw runs, as long as Playwright Chromium is available.
+
+| Platform | Status | Notes |
+|----------|--------|-------|
+| **macOS** (Mac Mini, MacBook — Intel & Apple Silicon) | ✅ Tested | Chromium installs automatically. Captions mode works great. No PulseAudio needed. |
+| **Linux x64** (Ubuntu, Debian, VPS, EC2) | ✅ Tested | May need system deps: `npx playwright-core install-deps chromium`. Headless servers need Xvfb. |
+| **Windows** (via WSL2) | ✅ Should work | Same as Linux x64. Not yet tested — feedback welcome. |
+| **Docker / Podman** | ✅ Should work | Run `npx playwright-core install-deps chromium` in container. Not yet tested. |
+| **Raspberry Pi** (ARM64, Pi 4/5) | ⚠️ Not tested | Playwright Chromium ARM builds can be unreliable. May need manual Chromium install. |
+| **Raspberry Pi 3** (ARMv7) | ❌ Not supported | Playwright does not ship ARMv7 Chromium. |
+
+### macOS (Mac Mini, MacBook)
+
+- `npx openbuilder` installs Chromium automatically
+- Audio capture mode is not available (no PulseAudio) — captions mode is the default and works great
+- Auth: `npx openbuilder auth` opens a browser window — sign in and press Enter
+- For headless/unattended operation: use `--auto` auth with `.env` credentials
+
+### Linux (Ubuntu, Debian, VPS, EC2)
+
+- Install system dependencies: `npx playwright-core install-deps chromium`
+- For headless servers (no display): start Xvfb first — `Xvfb :99 -screen 0 1280x720x24 &` and `export DISPLAY=:99`
+- Audio capture mode available if PulseAudio + ffmpeg are installed (optional — captions mode works without them)
+
+### VPS Providers
+
+Works on any VPS that runs OpenClaw:
+
+- **AWS** (EC2, Lightsail) — tested ✅
+- **Oracle Cloud** (Always Free tier) — should work
+- **Hetzner**, **Fly.io**, **GCP**, **Railway**, **Render** — should work (Linux x64)
+- **DigitalOcean** — should work
+
+Minimum: 1 vCPU, 1GB RAM (Chromium is the main resource consumer).
+
+### Requirements
+
+1. **Node.js 18+**
+2. **OpenClaw** running on the machine
+3. **A Google account** for the bot (or join as guest with `--anon`)
+4. **An AI API key** — Claude (Anthropic) or OpenAI — for meeting reports (optional but recommended)
+
+### Quick Setup (any platform)
+
+```bash
+npx openbuilder                           # Install skill + Chromium
+npx openbuilder auth                      # Sign into Google (one-time)
+npx openbuilder config set anthropicApiKey sk-ant-...  # For AI reports
+```
+
+Then tell your OpenClaw agent: "Join this meeting: https://meet.google.com/..."
+
+### Known Limitations
+
+- **Google Meet only** — Zoom and Teams support planned for a future release
+- **Captions depend on Google Meet's CC feature** — if Meet changes their DOM structure, caption scraping may break
+- **Audio capture mode is experimental** on headless servers — PulseAudio routing can be unreliable. Use `--captions` for reliability
+- **Bot appears as a participant** — other meeting participants will see "Super Liang" (or your bot's Google account name) in the People panel
+- **Google may block automated logins** — if Google flags the bot account, re-run `npx openbuilder auth` interactively or from a different IP
+- **One meeting at a time** per bot instance
+
 ## OpenClaw Integration
 
 OpenBuilder ships as an OpenClaw skill. After running `npx openbuilder`, it's available to your OpenClaw agent. The agent can:
 
 - Join meetings on your behalf
 - Capture and summarize transcripts
-- Generate full meeting reports
-- Send screenshots to your chat
+- Generate full meeting reports with action items and decisions
+- Send screenshots to your chat on demand
+- Tell you what's being discussed in real time
 
 See [SKILL.md](./SKILL.md) for the full agent integration guide.
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "openbuilder",
-  "version": "0.1.0",
+  "version": "0.1.2",
   "description": "Open-source AI meeting assistant — join Google Meet, capture transcripts, generate AI-powered meeting reports with summaries, action items, and speaker analytics",
   "type": "module",
   "bin": {

package/scripts/builder-auth.ts

@@ -9,7 +9,9 @@
  * Saves session to ~/.openbuilder/auth.json via Playwright's storageState.
  */
 
+import { execSync } from "node:child_process";
 import { existsSync, mkdirSync, readFileSync, writeFileSync } from "node:fs";
+import { homedir } from "node:os";
 import { createInterface } from "node:readline";
 import { config as dotenvConfig } from "dotenv";
 import { join, dirname } from "node:path";
@@ -163,15 +165,22 @@ async function main() {
     "--window-size=1280,720",
   ];
 
-  // Find full Chrome for headed mode
+  // Find full Chrome for headed mode (cross-platform)
   let executablePath: string | undefined;
-  const fullChromePath = join(
-    process.env.HOME || "~",
-    ".cache/ms-playwright/chromium-1208/chrome-linux64/chrome"
-  );
-  if (existsSync(fullChromePath)) {
-    executablePath = fullChromePath;
-  }
+  try {
+    const playwrightCache = join(homedir(), ".cache", "ms-playwright");
+    if (existsSync(playwrightCache)) {
+      const result = execSync(
+        process.platform === "darwin"
+          ? `find "${playwrightCache}" -path "*/Chromium.app/Contents/MacOS/Chromium" -type f 2>/dev/null | head -1`
+          : `find "${playwrightCache}" -path "*/chrome-linux*/chrome" -type f 2>/dev/null | head -1`,
+        { encoding: "utf-8" },
+      ).trim();
+      if (result && existsSync(result)) {
+        executablePath = result;
+      }
+    }
+  } catch { /* fallback to default */ }
 
   const browser = await pw.chromium.launch({
     headless,

package/scripts/builder-join.ts

@@ -529,6 +529,32 @@ async function waitForMeetingEnd(
       return "Navigated away from meeting";
     }
 
+    // Detect if redirected to Meet homepage (meeting ended and Google kicked us out)
+    const currentUrl = page.url();
+    if (
+      currentUrl === "https://meet.google.com/" ||
+      currentUrl === "https://meet.google.com" ||
+      currentUrl.match(/^https:\/\/meet\.google\.com\/?\?/) ||
+      currentUrl === "https://meet.google.com/landing"
+    ) {
+      return "Meeting ended (redirected to homepage)";
+    }
+
+    // Check if the Leave call button is gone (meeting UI disappeared)
+    try {
+      const leaveBtn = page.locator('[aria-label*="Leave call" i]').first();
+      const hasMeetingUI = await leaveBtn.isVisible({ timeout: 500 });
+      if (!hasMeetingUI) {
+        // Double-check: no meeting controls means we're not in a meeting
+        const hasAnyControls = await page.locator('[aria-label*="microphone" i], [aria-label*="camera" i]').first().isVisible({ timeout: 500 }).catch(() => false);
+        if (!hasAnyControls) {
+          return "Meeting ended (meeting UI gone)";
+        }
+      }
+    } catch {
+      // Couldn't check — continue
+    }
+
     // Check if all other participants have left
     const participantCount = await getParticipantCount(page);
 
@@ -1333,11 +1359,16 @@ export async function joinMeeting(opts: {
     let fullChromePath: string | undefined;
     if (useFullChrome) {
       try {
-        const result = execSync(
-          'find ~/.cache/ms-playwright/chromium-*/chrome-linux64 -name "chrome" -type f | head -1',
-          { encoding: "utf-8" },
-        ).trim();
-        fullChromePath = result || undefined;
+        const playwrightCache = join(homedir(), ".cache", "ms-playwright");
+        if (existsSync(playwrightCache)) {
+          const result = execSync(
+            process.platform === "darwin"
+              ? `find "${playwrightCache}" -path "*/Chromium.app/Contents/MacOS/Chromium" -type f 2>/dev/null | head -1`
+              : `find "${playwrightCache}" -path "*/chrome-linux*/chrome" -type f 2>/dev/null | head -1`,
+            { encoding: "utf-8" },
+          ).trim();
+          fullChromePath = result || undefined;
+        }
       } catch { /* fallback to default */ }
     }
     if (useFullChrome && fullChromePath) {
@@ -1368,8 +1399,12 @@ export async function joinMeeting(opts: {
     const fullChromePath = useFullChrome
       ? (() => {
           try {
+            const playwrightCache = join(homedir(), ".cache", "ms-playwright");
+            if (!existsSync(playwrightCache)) return undefined;
             const result = execSync(
-              'find ~/.cache/ms-playwright/chromium-*/chrome-linux64 -name "chrome" -type f | head -1',
+              process.platform === "darwin"
+                ? `find "${playwrightCache}" -path "*/Chromium.app/Contents/MacOS/Chromium" -type f 2>/dev/null | head -1`
+                : `find "${playwrightCache}" -path "*/chrome-linux*/chrome" -type f 2>/dev/null | head -1`,
               { encoding: "utf-8" },
             ).trim();
             return result || undefined;