Skip to content

Mingbird v2.1.3

Choose a tag to compare

View all tags
@Gustor-Wang Gustor-Wang released this 09 Oct 15:12
· 8 commits to main since this release
v2.1.3
5bb3283

鸣鸟 Mingbird v2.1.3

🇨🇳 中文在上 / English below.

让模型知道自己的工作目录——此前它只能从系统提示里看到主目录，常把主目录当成工作区。测试 560 项全绿（552 + 8 新增）。

修复

  • 现在每次任务都会告诉模型真实工作目录：静态系统提示里唯一的绝对路径是主目录（为"保存到桌面"而注入），工作目录本身从未告知模型——小模型直接把主目录当成工作区，选定了别的工作目录也复读主目录并拒绝越界。修复后系统提示动态追加「当前工作目录（所有命令与相对路径都在它之下运行）：<绝对路径>」，问答模式同样注入；静态 prefill 一个字节没动。
  • 工作目录随会话保存：会话元数据记录其工作目录，载入历史对话时自动恢复；新对话仍重置回默认目录（工作目录属于对话，不属于全局偏好）。

新增

  • 上下文档位补 512K 与 1M：新本地模型原生支持更大上下文，设置档位扩为 16K/32K/64K/128K/256K/512K/1M 七档。

English

Makes the model aware of its actual workspace — previously the only absolute path it could see was the home directory, which it often adopted as the workspace. Tests: 560 green (552 + 8 new).

Fixes

  • The model is now told its actual workspace on every task: the only absolute path in the static system prompt was the home directory (injected so models can save to Desktop etc.) — the workspace itself was never stated, so small models adopted the home directory as their workspace and refused to leave it. The system prompt now appends "current working directory (all commands and relative paths run under it): " dynamically per task; chat mode too. The static prefill is untouched, byte for byte.
  • The working directory is bound to the conversation: session metadata records its workdir and the GUI restores it when you load a history conversation; a new chat still resets to the default directory.

New

  • Context-window tiers up to 512K and 1M: the settings ladder is now 16K/32K/64K/128K/256K/512K/1M.

Install / 安装

Windows：下载 Mingbird-v2.1.3-CN-Setup.exe（中文）或 Mingbird-v2.1.3-EN-Setup.exe（English），双击安装。需要 Ollama 或任何 OpenAI 兼容本地服务；2~9B 小模型核显即可跑，全程数据不出本机。
Windows (EN): download Mingbird-v2.1.3-CN-Setup.exe (Chinese UI) or Mingbird-v2.1.3-EN-Setup.exe (English UI) and run the installer. Requires Ollama or any OpenAI-compatible local server; 2–9B models run on integrated graphics, and your data never leaves the machine.

macOS / Linux（实验性）：下载 mingbird-v2.1.3-macos-arm64.tar.gz / mingbird-v2.1.3-macos-x64.tar.gz / mingbird-v2.1.3-linux-x64.tar.gz，解压后运行 LocalAgent。系统语言自动识别（中文系统直接显示中文；英文系统输出自动跟随英文。也可在可执行文件旁建 app_lang.txt 写 zh/en 强制）。
macOS / Linux (experimental, EN): download the tar.gz for your arch, extract, run LocalAgent. System language is auto-detected (English systems get English output); you can also force it by placing an app_lang.txt (zh/en) next to the executable.

From source：pip install -r requirements.txt → python agent_gui.py（Python 3.12 推荐）。 / From source (EN): pip install -r requirements.txt → python agent_gui.py (Python 3.12 recommended).

Assets 7
Loading