AGENT SKILL

video-sdk/linux

Zoom Video SDK for Linux - C++ headless bots, raw audio/video capture/injection, Qt/GTK integration, Docker support

使用场景图像与多媒体
License 许可
Apache-2.0
最近更新
2026-11-13
Stars
★ 3.6w
Forks
3.9k

01 前言 Front Matter

name
video-sdk/linux
triggers
[ "linux video sdk", "zoom linux", "raw data linux", "qt zoom", "gtk zoom", "virtual audio linux" ]
description
Zoom Video SDK for Linux - C++ headless bots, raw audio/video capture/injection, Qt/GTK integration, Docker support
user-invocable
false

02 资源文件 20

03 SKILL.md 正文

Zoom Video SDK - Linux Development

Expert guidance for developing with the Zoom Video SDK on Linux. Build headless bots, raw media capture/injection applications, and custom UI integrations with Qt/GTK.

Official Documentation: https://developers.zoom.us/docs/video-sdk/linux/ API Reference: https://marketplacefront.zoom.us/sdk/custom/linux/ Sample Repository: https://github.com/zoom/videosdk-linux-raw-recording-sample

Quick Links

New to Video SDK? Follow this path:

  1. SDK Architecture Pattern - Universal 3-step pattern for ANY feature
  2. Session Join Pattern - Complete working code to join a session
  3. Raw Data vs Canvas - CRITICAL: Linux has NO Canvas API - raw data ONLY
  4. Raw Video Capture - Capture and process YUV420 frames

Reference:

Having issues?

Key Differences from Windows/macOS

Feature Linux Windows/Mac
Canvas API ❌ Not available ✅ Available
Raw Data Pipe ✅ ONLY option ✅ Available
UI Integration Qt, GTK, SDL2, OpenGL Win32/WinForms/WPF, Cocoa
Headless Support ✅ Excellent (Docker) Limited
Audio PulseAudio required Native
Virtual Devices ✅ Required for headless Optional

SDK Overview

The Zoom Video SDK for Linux is a C++ library optimized for:

  • Headless Bots: Docker/WSL support, no display required
  • Raw Data Access: Capture YUV420 video, PCM audio
  • Raw Data Injection: Virtual camera/mic for custom media
  • Screen Sharing: Capture or inject share data
  • Cloud Recording: Record sessions to Zoom cloud
  • Live Streaming: Stream to RTMP endpoints
  • Live Transcription: Real-time speech-to-text
  • Qt/GTK Integration: Full UI framework support

Prerequisites

System Requirements

  • OS: Ubuntu 20.04+, Debian 11+, or compatible
  • Architecture: x64 (recommended), ARM64
  • Compiler: GCC 9+, Clang 10+
  • CMake: 3.14 or later
  • Qt5: Bundled with SDK (do NOT install system Qt5)

Dependencies

sudo apt update
sudo apt install -y build-essential gcc cmake libglib2.0-dev liblzma-dev \
    libxcb-image0 libxcb-keysyms1 libxcb-xfixes0 libxcb-xkb1 libxcb-shape0 \
    libxcb-shm0 libxcb-randr0 libxcb-xtest0 libgbm1 libxtst6 libgl1 libnss3 \
    libasound2 libpulse0

# For headless Linux
sudo apt install -y pulseaudio

# PulseAudio configuration (CRITICAL for audio)
mkdir -p ~/.config
echo "[General]" > ~/.config/zoomus.conf
echo "system.audio.type=default" >> ~/.config/zoomus.conf

# Log directory
mkdir -p ~/.zoom/logs

Quick Start

#include "zoom_video_sdk_api.h"
#include "zoom_video_sdk_interface.h"
#include "zoom_video_sdk_delegate_interface.h"

USING_ZOOM_VIDEO_SDK_NAMESPACE

// 1. Create SDK
IZoomVideoSDK* sdk = CreateZoomVideoSDKObj();

// 2. Initialize
ZoomVideoSDKInitParams init_params;
init_params.domain = "https://zoom.us";
init_params.enableLog = true;
init_params.logFilePrefix = "bot";
init_params.videoRawDataMemoryMode = ZoomVideoSDKRawDataMemoryModeHeap;
init_params.shareRawDataMemoryMode = ZoomVideoSDKRawDataMemoryModeHeap;
init_params.audioRawDataMemoryMode = ZoomVideoSDKRawDataMemoryModeHeap;

sdk->initialize(init_params);

// 3. Add delegate
sdk->addListener(myDelegate);

// 4. Join session
ZoomVideoSDKSessionContext ctx;
ctx.sessionName = "my-session";
ctx.userName = "Linux Bot";
ctx.token = "jwt-token";
ctx.audioOption.connect = true;
ctx.audioOption.mute = false;
ctx.videoOption.localVideoOn = false;

// For headless: Virtual audio speaker
ctx.virtualAudioSpeaker = new VirtualSpeaker();

IZoomVideoSDKSession* session = sdk->joinSession(ctx);

See Session Join Pattern for complete code.

Key Features

Feature Linux Support Guide
Session Management ✅ Full Session Join
Raw Video (YUV420) ✅ ONLY rendering option Raw Video
Raw Audio (PCM) ✅ Full Raw Audio
Virtual Camera/Mic ✅ Full Virtual Devices
Cloud Recording ✅ Full Recording
Live Streaming ✅ Full Live Stream
Live Transcription ✅ Full Transcription
Command Channel ✅ Full Commands
Chat ✅ Full Chat
Qt Integration ✅ Recommended Qt/GTK
GTK Integration ✅ Supported Qt/GTK
Docker/Headless ✅ Excellent Virtual Devices

Critical Gotchas

⚠️ CRITICAL #1: No Canvas API on Linux

Problem: Linux SDK does NOT have Canvas API like Windows/Mac.

Solution: You MUST use Raw Data Pipe and implement your own rendering.

See: Raw Data vs Canvas

⚠️ CRITICAL #2: PulseAudio Required for Audio

Problem: SDK requires PulseAudio for raw audio functions.

Solution:

sudo apt install -y pulseaudio
mkdir -p ~/.config
echo "[General]" > ~/.config/zoomus.conf
echo "system.audio.type=default" >> ~/.config/zoomus.conf

See: PulseAudio Setup

⚠️ CRITICAL #3: Qt5 Dependencies

Problem: SDK requires Qt5 libraries (bundled, NOT system Qt5).

Solution:

# Copy from SDK package
cp -r samples/qt_libs/Qt/lib/* lib/zoom_video_sdk/

# Create symlinks
cd lib/zoom_video_sdk
for lib in libQt5*.so.5; do ln -sf $lib ${lib%.5}; done

See: Qt Dependencies

⚠️ CRITICAL #4: Heap Memory Mode

Always use heap mode for raw data:

init_params.videoRawDataMemoryMode = ZoomVideoSDKRawDataMemoryModeHeap;
init_params.shareRawDataMemoryMode = ZoomVideoSDKRawDataMemoryModeHeap;
init_params.audioRawDataMemoryMode = ZoomVideoSDKRawDataMemoryModeHeap;

⚠️ CRITICAL #5: Virtual Audio for Headless

Problem: Docker/headless environments have no audio devices.

Solution: Use virtual audio speaker and mic.

session_context.virtualAudioSpeaker = new VirtualSpeaker();
session_context.virtualAudioMic = new VirtualMic();

See: Virtual Audio/Video

Sample Repositories

Official Samples

Repository Description
raw-recording-sample Raw audio/video capture
qt-quickstart Qt6 UI integration
gtk-quickstart GTK3 UI integration

Sample Architecture

Headless Bot (Docker):
┌──────────────────────────────────┐
│  Virtual Audio Speaker/Mic       │
├──────────────────────────────────┤
│  Raw Data Processing             │
│  - YUV420 → File/Stream   


## Merged from video-sdk/linux/SKILL.md

# Zoom Video SDK Linux - Complete Documentation Index

## Quick Start Path

**If you're new to the SDK, follow this order:**

1. **Read the architecture pattern** → [concepts/sdk-architecture-pattern.md](concepts/sdk-architecture-pattern.md)
   - Universal formula: Singleton → Delegate → Subscribe
   - Once you understand this, you can implement any feature

2. **Understand Linux specifics** → [concepts/raw-data-vs-canvas.md](concepts/raw-data-vs-canvas.md)
   - **CRITICAL**: Linux has NO Canvas API - raw data ONLY

3. **Implement session join** → [examples/session-join-pattern.md](examples/session-join-pattern.md)
   - Complete working JWT + session join code

4. **Setup environment** → [troubleshooting/pulseaudio-setup.md](troubleshooting/pulseaudio-setup.md)
   - PulseAudio configuration (required for audio)
   - [troubleshooting/qt-dependencies.md](troubleshooting/qt-dependencies.md)
   - Qt5 library setup (bundled with SDK)

5. **Implement features** → Choose from examples below

---

## Documentation Structure

video-sdk/linux/ ├── SKILL.md # Main skill overview ├── SKILL.md # This file - navigation guide ├── linux.md # Platform summary │ ├── concepts/ # Core architectural patterns │ ├── sdk-architecture-pattern.md # Universal formula for ANY feature │ ├── singleton-hierarchy.md # 5-level navigation guide │ └── raw-data-vs-canvas.md # Linux-specific: raw data ONLY │ ├── examples/ # Complete working code │ ├── session-join-pattern.md # JWT auth + session join │ └── command-channel.md # Command channel with threading │ ├── troubleshooting/ # Problem solving guides │ ├── pulseaudio-setup.md # Audio configuration │ ├── qt-dependencies.md # Qt5 library setup │ ├── build-errors.md # Common build issues │ └── common-issues.md # Quick diagnostic workflow │ └── references/ # Reference documentation └── linux-reference.md # API hierarchy, methods, error codes


---

## By Use Case

### I want to build a headless bot
1. [SDK Architecture Pattern](concepts/sdk-architecture-pattern.md) - Understand the pattern
2. [Session Join Pattern](examples/session-join-pattern.md) - Join sessions
3. [PulseAudio Setup](troubleshooting/pulseaudio-setup.md) - Configure audio
4. [Raw Data vs Canvas](concepts/raw-data-vs-canvas.md) - Understand Linux differences

### I'm getting build errors
1. [Build Errors Guide](troubleshooting/build-errors.md) - SDK build issues
2. [Qt Dependencies](troubleshooting/qt-dependencies.md) - Qt5 setup
3. [Common Issues](troubleshooting/common-issues.md) - Quick diagnostics

### I'm getting runtime errors
1. [PulseAudio Setup](troubleshooting/pulseaudio-setup.md) - Audio not working
2. [Qt Dependencies](troubleshooting/qt-dependencies.md) - Library not found
3. [Common Issues](troubleshooting/common-issues.md) - Error code tables

### I want to use command channel
1. [Command Channel](examples/command-channel.md) - Send/receive commands
2. [Common Issues](troubleshooting/common-issues.md) - Threading requirements

### I want to implement a specific feature
1. [SDK Architecture Pattern](concepts/sdk-architecture-pattern.md) - **START HERE!**
2. [Singleton Hierarchy](concepts/singleton-hierarchy.md) - Navigate to the feature
3. [API Reference](references/linux-reference.md) - Method signatures

---

## Most Critical Documents

### 1. SDK Architecture Pattern (MASTER DOCUMENT)
**[concepts/sdk-architecture-pattern.md](concepts/sdk-architecture-pattern.md)**

The universal 3-step pattern:
1. Get singleton (SDK, helpers, session, users)
2. Implement delegate (event callbacks)
3. Subscribe and use

### 2. Raw Data vs Canvas (LINUX-SPECIFIC)
**[concepts/raw-data-vs-canvas.md](concepts/raw-data-vs-canvas.md)**

**CRITICAL**: Unlike Windows/Mac, Linux SDK has NO Canvas API. You MUST use raw data pipe.

### 3. PulseAudio Setup (MOST COMMON ISSUE)
**[troubleshooting/pulseaudio-setup.md](troubleshooting/pulseaudio-setup.md)**

Audio requires PulseAudio configuration.

### 4. Qt Dependencies
**[troubleshooting/qt-dependencies.md](troubleshooting/qt-dependencies.md)**

SDK requires bundled Qt5 libraries, NOT system Qt5.

---

## Key Learnings

### Critical Discoveries:

1. **Linux has NO Canvas API**
   - Windows/Mac have Canvas API for SDK-rendered video
   - Linux MUST use Raw Data Pipe
   - See: [Raw Data vs Canvas](concepts/raw-data-vs-canvas.md)

2. **PulseAudio is MANDATORY**
   - SDK requires PulseAudio for raw audio
   - Must configure ~/.config/zoomus.conf
   - See: [PulseAudio Setup](troubleshooting/pulseaudio-setup.md)

3. **Use Bundled Qt5, NOT System Qt5**
   - SDK includes specific Qt5 versions
   - Copy from samples/qt_libs/
   - See: [Qt Dependencies](troubleshooting/qt-dependencies.md)

4. **Helpers Control YOUR Streams Only**
   - `videoHelper->startVideo()` starts YOUR camera
   - To see others, subscribe to their VideoPipe
   - See: [Singleton Hierarchy](concepts/singleton-hierarchy.md)

5. **Virtual Devices for Headless**
   - Docker/headless needs virtual audio speaker/mic
   - Set before joining session
   - See: [Session Join Pattern](examples/session-join-pattern.md)

6. **Always Use Heap Memory Mode**
   ```cpp
   init_params.videoRawDataMemoryMode = ZoomVideoSDKRawDataMemoryModeHeap;
  1. GLib Main Loop Required

    • while/sleep loops don't dispatch SDK events
    • Must use g_main_loop_run()
    • See: Common Issues
  2. All SDK Calls Must Be on Main Thread

    • Background thread SDK calls return error 2 (Internal_Error)
    • Use g_idle_add() to schedule on GLib main thread
    • See: Command Channel
  3. Command Channel is Session-Scoped

    • Does NOT span across different sessions
    • Both sender and receiver must be in the same session
    • See: Command Channel

Sample Repositories


Quick Reference

"My code won't compile"

→ Build Errors Guide

"Audio not working"

→ PulseAudio Setup

"Library not found"

→ Qt Dependencies

"How do I implement [feature]?"

→ SDK Architecture Pattern

"What error code means what?"

→ Common Issues


Document Version

Based on Zoom Video SDK for Linux v2.x


Happy coding!

Remember: The SDK Architecture Pattern is your key to unlocking the entire SDK. Read it first!

Operations

  • RUNBOOK.md - 5-minute preflight and debugging checklist.

04 同仓库更多 Skills

anthropics
anthropics/knowledge-work-plugins

ticket-deflector

Reads a forwarded customer email or ticket, pulls order and refund status from a payments connector (PayPal, Square, or Stripe) or Shopify, account history from the CRM, and open tickets from a support desk (Zoho Desk), drafts a tone-matched reply in the owner's writing voice, and can issue a refund through the payments connector with explicit owner approval. With Shopify connected it also runs a proactive order-triage mode that surfaces orders needing attention — unfulfilled past the promised window, payment problems, pending refunds, stuck shipments — and drafts the next action for each before the customer has to ask. Use when the user says "draft a response," "answer this customer," "where's my order," "I want a refund," "check my orders," or "anything about to blow up."

★ 3.6w
anthropics
anthropics/knowledge-work-plugins

tax-season-organizer

Prepares tax-season materials for the owner's accountant, not tax advice. US federal tax; a non-US business gets its closed-books packet instead. Two modes: (1) quarterly estimated tax from YTD net income in the ledger (MYOB, NetSuite, QuickBooks, Xero, or Zoho Books); (2) year-end 1099 prep, scanning the ledger, PayPal, and Stripe for contractors paid over USD 600 into a 1099-NEC list with missing W-9 flags. Any tax request routes first to /tax-prep, which confirms the books are closed and reconciled before running this skill. Use this skill directly only when the owner says the period's books are already closed: "books are closed, now do the 1099s," "run the quarterly estimate off the closed numbers," or "just the contractor W-9 list."

★ 3.6w
anthropics
anthropics/knowledge-work-plugins

tax-prep

Prepares tax materials as a two-link chain — month-end-prep confirms the books are closed and reconciled first, then tax-season-organizer calculates the quarterly estimated payment or builds the year-end 1099-NEC list and accountant packet from those closed numbers. Requires a ledger (MYOB, NetSuite, QuickBooks, Xero, or Zoho Books); uses Gusto, PayPal, and Stripe when connected, else CSV upload. US federal tax math; a non-US business gets the closed-books packet. Prep material for a CPA, never tax advice, nothing filed. Trigger on "quarterly taxes," "estimated tax payment," "how much should I set aside," "1099s," "1099-NEC," "W-9s," "year-end tax prep," "get my books ready for my accountant," any phrasing that suggests a tax deadline is coming, or a question about net profit or YTD income that sounds like worry about a tax bill. An owner who says the period's books are already closed routes to tax-season-organizer directly.

★ 3.6w
anthropics
anthropics/knowledge-work-plugins

speed-to-lead

Makes sure no inbound lead goes unanswered. Watches web forms, the shared inbox, and the CRM for new inquiries, qualifies each one against the owner's criteria, drafts a reply in the owner's voice with real meeting times for the owner to approve and send, routes anything hot to a human right away, and logs the whole thing to the CRM. Never sends on its own. Built to run on a schedule so overnight inquiries are drafted and sorted by morning, and works from forwarded or pasted inquiries when no connector is available. Use this whenever the owner is worried about inbound leads or response time — including phrasings like "leads are going cold," "did anyone answer that inquiry," "I keep missing form submissions," "nothing falls through the cracks," "handle my inbound," "respond to this lead," or "we lose people because nobody gets back fast enough." Reach for it for a single unanswered inquiry as readily as for a standing setup.

行政与办公人员
★ 3.6w
anthropics
anthropics/knowledge-work-plugins

social-content-engine

Runs the owner's content operation as a standing thing rather than a one-off: keeps a rolling posting calendar, learns the owner's voice, generates on-brand Canva graphics, writes captions that sound like them, repurposes one asset into every channel format, and stages posts for approval instead of publishing on its own. Pulls product photos and prices from Shopify when connected; with no connectors, drafts the calendar and asset briefs for manual posting. Use this whenever the owner talks about posting, content, or social — including "I need to post more," "we've gone quiet on Instagram," "what should we post this month," "make the content," "give me a month of posts," "write this in my voice," "repurpose this," or "set up our content calendar." Reach for it for a single post as readily as for a full quarter. One handoff: a finished, approved campaign brief goes to canva-creator; this skill owns the standing operation.

内容创作者
★ 3.6w
anthropics
anthropics/knowledge-work-plugins

smb-router

The front door to the Small Business plugin. Listens to what the owner needs right now — vague or specific — and routes them to the best skill or slash command for the moment. Also serves as a guide: explains what's available, suggests what to try next, and adapts recommendations based on stored business context. Trigger whenever the owner asks "what can you do," "help me with my business," "what should I focus on," "I don't know where to start," or any open-ended business request that doesn't clearly match a single skill.

★ 3.6w