Agenvoy v1.0.0 Release Notes

v0.35.5 -> v1.0.0

Agenvoy 1.0.0

Agenvoy 1.0.0 is the first formal release of a Taiwan-developed AI Agent Harness that turns conversation into completed work on your computer. It coordinates models, context, tools, real-time data, task routing, memory, schedules, and execution in one local workflow, while you retain control of your files and environment.

翻譯

Agenvoy 1.0.0 是台灣自研 AI Agent Harness 的首個正式版本，將對話轉化為在你的電腦上實際完成的工作。它在本機工作流程中協調模型、工作脈絡、工具、即時資料、任務路由、記憶、排程與執行，同時讓檔案與環境的控制權留在你手上。

What Agenvoy 1.0.0 delivers

Live research and tool-driven work — Retrieve current information, work with local files, and complete multi-step tasks. When a suitable capability is missing, Agenvoy can create, test, and retain a reusable tool.
Local file search and working context — Search indexed local files and use their contents as task context without moving the work outside your environment.
Automation in plain language — Turn one-sentence requests into scheduled work that runs locally and delivers results through configured channels.
Shared sandboxed tools through MCP — Connect Claude Code, Codex, OpenCode, and other MCP clients to one shared tool library instead of rebuilding the same integrations per agent.
Visible, controlled execution — Review command output in the TUI or Web dashboard; sensitive paths and restricted actions continue to require confirmation.
Model and service integrations — Route work across configured models and connect image generation, speech-to-text, text-to-speech, stdio/HTTP MCP servers, and OAuth services.
Private remote access — Use the Web interface, Telegram, or Discord while the local daemon establishes outbound connections and does not require public inbound ports.
Hands-free Web voice interaction — Speak naturally, use wake-word detection, and interrupt spoken replies while preserving the full result in chat.

翻譯

即時研究與工具執行 — 取得即時資訊、處理本機檔案並完成多步驟任務；缺少合適能力時，Agenvoy 可建立、測試並保留可重用的工具。
本機檔案搜尋與工作脈絡 — 搜尋已建立索引的本機檔案，直接把內容作為任務脈絡，不必將工作移出你的環境。
自然語言自動化 — 一句話建立排程工作，在本機執行並透過已設定的管道交付結果。
透過 MCP 共用沙箱工具 — 讓 Claude Code、Codex、OpenCode 與其他 MCP 用戶端共用同一套工具庫，無須為每個 Agent 重複建立整合。
可見且可控的執行流程 — 在 TUI 或 Web 儀表板查看命令輸出；敏感路徑與受限操作仍須確認。
模型與服務整合 — 在已設定的模型間路由工作，並串接圖片生成、語音辨識、文字轉語音、stdio／HTTP MCP server 與 OAuth 服務。
私密的遠端存取 — 透過 Web、Telegram 或 Discord 使用 Agenvoy；本機 daemon 主動向外連線，不需要公開入站連接埠。
免持 Web 語音互動 — 自然說話、使用喚醒詞並可打斷語音回覆，同時將完整結果保留在聊天內容中。

Change the project license from Apache-2.0 to AGPL-3.0

The project license is now AGPL-3.0.

Migration:

Review distribution, modification, and network-use obligations under AGPL-3.0 before upgrading or redistributing the project.
Replace bundled Apache-2.0 notices with the current project license where applicable.
Keep third-party notices and their original license terms intact.

翻譯

專案授權條款已由 Apache-2.0 改為 AGPL-3.0。

遷移方式：

升級或重新散布專案前，檢查 AGPL-3.0 對散布、修改與網路使用的要求。
在適用範圍內，將隨附的 Apache-2.0 聲明替換為目前專案授權條款。
保留第三方元件及其原有授權條款與聲明。

Release scope

First formal release of Agenvoy as a Taiwan-developed AI Agent Harness.
LICENSE — project license changed to AGPL-3.0.

Generated by SKILL