Release date: 2026.6.10
Downloads
| Platform | File | Size |
|---|---|---|
| macOS (arm64) | PilotDeck-0.1.0-mac-arm64.dmg | 178 MB |
| Windows (x64) | PilotDeck-0.1.0-win-x64.exe | 160 MB |
PilotDeck Desktop v0.1.0
PilotDeck now ships native desktop clients for Windows and macOS. Download, install, and start using PilotDeck locally — no manual deployment needed.
This release also brings concentrated stability improvements (38 fixes), a refreshed Settings UI, IM channel enhancements, and new model provider support.
Highlights
Desktop Clients
- Native Windows & macOS desktop apps — download and run, zero deployment overhead.
Stability (38 fixes)
- Large-file I/O hardened with auto-recovery & circuit-breaker protection; no more stalls or data loss on interrupted writes.
- Streaming UI updates optimized — noticeably lower CPU usage during long sessions.
Settings Overhaul
- Simplified layout, beginner-friendly navigation.
- In-app version check & one-click upgrade.
- Cron task configuration via GUI.
- MCP server management directly in Settings — no more hand-editing YAML.
New Model Provider
- Volcengine (火山引擎) models now supported out of the box.
System Requirements
- macOS 13+ / Windows 10+
- 16 GB+ RAM recommended for on-device model inference
Known Limitations
- Some file operations on Windows may hit EPERM issues (partially mitigated).
Links
PilotDeck 桌面版 v0.1.0
PilotDeck 正式推出 Windows 和 macOS 桌面客户端，下载即用，无需手动部署。
本次更新同时集中修复了 38 个稳定性问题，重做了设置界面，新增 IM 渠道能力与模型接入。
下载
| 平台 | 文件 | 大小 |
|---|---|---|
| macOS (arm64) | PilotDeck-0.1.0-mac-arm64.dmg | 178 MB |
| Windows (x64) | PilotDeck-0.1.0-win-x64.exe | 160 MB |
亮点
桌面客户端
- Windows & macOS 原生桌面应用，开箱即用，本地化体验拉满。
稳定性（38 项修复）
- 大文件读写全面加固：新增截断自动恢复 + 断路器保护，大文件操作不再卡死或丢数据。
- UI 流式更新逻辑优化，CPU 占用大幅降低，长时间使用不再卡顿。
Settings 焕新
- 界面全面简化，新手也能快速找到目标选项。
- 新增版本检查与一键升级。
- Cron 定时任务可视化配置。
- MCP 服务器直接在界面中添加管理，不再需要手动改配置文件。
新增模型支持
- 火山引擎（字节跳动）模型开箱即用。
系统要求
- macOS 13+ / Windows 10+
- 推荐 16GB+ 内存（端侧模型推理时）
已知限制
- Windows 下部分文件操作存在 EPERM 兼容性问题（已部分修复）。
相关链接
PilotDeck is jointly developed by Tsinghua THUNLP, ModelBest, OpenBMB, and AI9Stars.