AISRT v0.1.5

AISRT v0.1.5

简体中文

本次 Safety Hotfix 重点保护批量字幕输出和已有用户文件。AISRT 现在会在加载模型前检查批内输出冲突，严格拒绝无效 SRT，并通过原子替换写入字幕；模型权重、PyTorch/CUDA、FFmpeg 和运行缓存仍不随包分发。

主要变化

• 批量任务在模型加载前统一规划最终字幕、翻译字幕和预览字幕；同名目标会直接阻止开始，不再静默覆盖。
• 不再按媒体文件名删除同名 .txt 或 .raw.srt，用户笔记和其他工具生成的文件保持不变。
• 最终字幕、翻译字幕和预览字幕改为原子替换写入；替换失败时保留已有文件。
• 无效、部分损坏或空 SRT 会明确失败且不写输出；翻译输入与输出路径相同时也会拒绝执行。
• 新增覆盖上述数据安全边界的回归测试，并同步更新中英文说明。

发布资产

• AiSRT-v0.1.5-windows-portable.zip：轻量 Windows 包，包含源码、AISRT.bat 一键启动脚本和依赖文件。
• aisrt-0.1.5-py3-none-any.whl：Python wheel，适合手动安装的用户。
• SHA256SUMS.txt：发布资产 SHA256 校验值。

注意事项

• 为控制安装包大小，Python、PyTorch/CUDA、模型权重和 FFmpeg 不随 ZIP 分发。
• 解压 ZIP 后优先双击 AISRT.bat。首次启动会自动调用 install_runtime.bat 创建 .venv 并安装 Python 依赖，后续会直接打开 GUI。
• 使用远程模型 ID 时，AISRT 会在首次运行时下载模型权重到配置的 Hugging Face 缓存目录。
• FFmpeg 和 ffprobe 需要单独安装，并确保可在 PATH 中直接运行。

English

This Safety Hotfix protects batch subtitle output and existing user files. AISRT now checks batch-internal output collisions before loading models, strictly rejects invalid SRT input, and writes subtitles through atomic replacement. Model weights, PyTorch/CUDA, FFmpeg, and runtime caches remain unbundled.

Highlights

• Plans final, translated, and preview subtitle targets before model loading; duplicate targets stop the batch instead of being silently overwritten.
• No longer deletes same-stem .txt or .raw.srt files that may belong to users or other tools.
• Writes final, translated, and preview subtitles through atomic replacement so failed writes preserve existing files.
• Rejects invalid, partially malformed, or empty SRT input without writing output, and refuses translation when input and output paths are the same.
• Adds regression coverage for these data-safety boundaries and updates the bilingual documentation.

Assets

• AiSRT-v0.1.5-windows-portable.zip: lightweight Windows package with source code, the one-click AISRT.bat launcher and dependency files.
• aisrt-0.1.5-py3-none-any.whl: Python wheel for users who prefer manual Python installation.
• SHA256SUMS.txt: SHA256 checksums for release assets.

Notes

• To keep the download small, Python, PyTorch/CUDA, model weights and FFmpeg are not bundled in the ZIP.
• After extracting the ZIP, double-click AISRT.bat first. The first launch calls install_runtime.bat automatically to create .venv and install Python dependencies; later launches open the GUI directly.
• When remote model IDs are used, AISRT downloads model weights to the configured Hugging Face cache on first use.
• FFmpeg and ffprobe must be installed separately and available in PATH.

AISRT v0.1.4

AISRT v0.1.4

简体中文

本次发布重点降低普通用户的首次使用门槛，同时保持 Release 包足够轻量。Windows portable ZIP 现在提供 AISRT.bat 一键启动入口，首次启动会自动安装 Python 依赖到本地 .venv，后续直接打开 GUI；模型权重、PyTorch/CUDA、FFmpeg 和运行缓存仍不随包分发，避免下载体积失控。

主要变化

• 新增识别过程中的 movie.preview.srt 预览字幕，长视频处理时可以先查看已完成片段；完整处理成功后会删除预览并重写最终 movie.srt。
• Windows portable ZIP 新增 AISRT.bat，普通用户解压后优先双击它；start_gui.bat 和 open_shell.bat 保留给手动启动和排查。
• install_runtime.bat 会清晰报告依赖安装失败；环境检查发现 FFmpeg 等问题时给出警告，但不阻止 GUI 打开。
• README 和英文 README 重新整理首屏介绍、快速入口、SEO 关键词、预览字幕说明和发布包说明。
• 刷新中英文 GUI 截图，移除旧截图里的本机路径。
• 打包规则新增一键启动器、英文 README 和 forbidden payload 检查，继续排除模型、运行时、缓存、媒体、字幕和日志。
• GUI 图标资源改为集中打包到 aisrt/assets/icons/，并补充对应测试。

发布资产

• AiSRT-v0.1.4-windows-portable.zip：轻量 Windows 包，包含源码、AISRT.bat 一键启动脚本和依赖文件。
• aisrt-0.1.4-py3-none-any.whl：Python wheel，适合手动安装的用户。
• SHA256SUMS.txt：发布资产 SHA256 校验值。

注意事项

• 为控制安装包大小，Python、PyTorch/CUDA、模型权重和 FFmpeg 不随 ZIP 分发。
• 解压 ZIP 后优先双击 AISRT.bat。首次启动会自动调用 install_runtime.bat 创建 .venv 并安装 Python 依赖，后续会直接打开 GUI。
• 使用远程模型 ID 时，AISRT 会在首次运行时下载模型权重到配置的 Hugging Face 缓存目录。
• FFmpeg 和 ffprobe 需要单独安装，并确保可在 PATH 中直接运行。

English

This release focuses on lowering the first-run barrier for regular users while keeping the release package small. The Windows portable ZIP now includes a one-click AISRT.bat launcher. On first launch it installs Python dependencies into the local .venv; later launches open the GUI directly. Model weights, PyTorch/CUDA, FFmpeg and runtime caches are still not bundled, so the download stays lightweight.

Highlights

• Added movie.preview.srt during ASR recognition so long jobs can be inspected before final post-processing rewrites movie.srt.
• Added AISRT.bat to the Windows portable ZIP for the primary user path; start_gui.bat and open_shell.bat remain for manual startup and troubleshooting.
• Improved install_runtime.bat failure handling for dependency installation; environment issues such as missing FFmpeg are shown as warnings without blocking the GUI.
• Reworked the Chinese and English README files for first-time users, search discoverability, preview subtitles and release-package expectations.
• Refreshed Chinese and English GUI screenshots without local private paths.
• Hardened portable packaging checks for the launcher, bilingual README files and forbidden payloads while continuing to exclude models, runtimes, caches, media, subtitles and logs.
• Packaged the refreshed GUI SVG icon set under aisrt/assets/icons/ and added matching tests.

Assets

• AiSRT-v0.1.4-windows-portable.zip: lightweight Windows package with source code, the one-click AISRT.bat launcher and dependency files.
• aisrt-0.1.4-py3-none-any.whl: Python wheel for users who prefer manual Python installation.
• SHA256SUMS.txt: SHA256 checksums for release assets.

Notes

• To keep the download small, Python, PyTorch/CUDA, model weights and FFmpeg are not bundled in the ZIP.
• After extracting the ZIP, double-click AISRT.bat first. The first launch calls install_runtime.bat automatically to create .venv and install Python dependencies; later launches open the GUI directly.
• When remote model IDs are used, AISRT downloads model weights to the configured Hugging Face cache on first use.
• FFmpeg and ffprobe must be installed separately and available in PATH.

AISRT v0.1.3

AISRT v0.1.3

简体中文

本次发布提供轻量 Windows portable 源码包、Python wheel，以及实验性的 Windows x64 CUDA 完整运行时安装包。主要变化仍是优化本地 ASR 默认参数，在不切换模型、不提高批大小的前提下改善常规处理速度。

主要变化

• 将 ASR 每个音频块的默认 max_new_tokens 从 2048 降到 1536，减少异常长生成带来的耗时。
• 保留稳定的默认组合：1.7B ASR 模型、45 秒音频分块和 batch size 1。
• CLI 帮助、GUI 推荐模式和底层模型加载默认值已统一使用新的生成上限。
• 中英文 README 增加 --max-new-tokens 说明，长对白被截断时仍可临时调高。
• 增加 CLI 和 GUI 默认参数测试，降低后续改动误改推荐档位的风险。

发布资产

• AiSRT-v0.1.3-windows-x64-cuda-runtime-setup.exe：实验性 Windows x64 CUDA 完整运行时安装包，包含 Python、PyQt6、PyTorch/CUDA 运行库和 AISRT 程序，不包含模型权重和 FFmpeg。
• AiSRT-v0.1.3-windows-portable.zip：轻量 Windows 包，包含源码、启动脚本和依赖文件。
• aisrt-0.1.3-py3-none-any.whl：Python wheel，适合手动安装的用户。
• SHA256SUMS.txt：发布资产 SHA256 校验值。

注意事项

• 完整运行时安装包不包含模型权重、模型缓存、FFmpeg、媒体文件、生成字幕、截图、测试文件或日志。
• 完整运行时安装包仍需要 NVIDIA CUDA 显卡和兼容的 NVIDIA 驱动；用户不需要单独安装 CUDA Toolkit。
• 使用远程模型 ID 时，AISRT 会在首次运行时下载模型权重到配置的 Hugging Face 缓存目录。
• portable ZIP 不包含 Python、PyTorch/CUDA 运行库、模型权重、模型缓存、FFmpeg、媒体文件、生成字幕、截图、测试文件或日志。
• 使用 portable ZIP 时，解压后运行 install_runtime.bat，脚本会创建 .venv 并安装 Python 依赖。
• FFmpeg 和 ffprobe 需要单独安装，并确保可在 PATH 中直接运行。

English

This release provides a lightweight Windows portable source package, a Python wheel, and an experimental Windows x64 CUDA full-runtime installer. The main change remains tuning local ASR defaults for better everyday speed without changing the model or raising batch size.

Highlights

• Lowered the default ASR max_new_tokens per audio chunk from 2048 to 1536 to reduce time spent on unusually long generations.
• Kept the stable default combination: 1.7B ASR model, 45-second audio chunks and batch size 1.
• Unified the CLI help, GUI recommended profile and model-loading default around the new generation limit.
• Documented --max-new-tokens in both README files; users can still raise it temporarily if long dialogue is truncated.
• Added CLI and GUI default-parameter tests to keep the recommended profile from drifting accidentally.

Assets

• AiSRT-v0.1.3-windows-x64-cuda-runtime-setup.exe: experimental Windows x64 CUDA full-runtime installer with Python, PyQt6, PyTorch/CUDA runtime libraries and AISRT, excluding model weights and FFmpeg.
• AiSRT-v0.1.3-windows-portable.zip: lightweight Windows package with source code, launch scripts and dependency files.
• aisrt-0.1.3-py3-none-any.whl: Python wheel for users who prefer manual Python installation.
• SHA256SUMS.txt: SHA256 checksums for release assets.

Notes

• The full-runtime installer does not include model weights, model caches, FFmpeg, media files, generated subtitles, screenshots, tests or logs.
• The full-runtime installer still requires an NVIDIA CUDA GPU and a compatible NVIDIA driver; users do not need to install the CUDA Toolkit separately.
• When remote model IDs are used, AISRT downloads model weights to the configured Hugging Face cache on first use.
• The portable ZIP does not include Python, PyTorch/CUDA runtime libraries, model weights, model caches, FFmpeg, media files, generated subtitles, screenshots, tests or logs.
• For the portable ZIP, run install_runtime.bat after extracting it to create .venv and install Python dependencies.
• FFmpeg and ffprobe must be installed separately and available in PATH.

AISRT v0.1.2

AISRT v0.1.2

简体中文

本次发布提供轻量 Windows portable 源码包和 Python wheel，重点改进本地 SRT 翻译、GUI 多语言体验和日志提示。

主要变化

• 新增本地 SRT 翻译能力，可在 GUI 中翻译已有 SRT，也可通过 ai-sub-translate 在命令行批处理。
• GUI 主流程支持“识别并翻译”，会先保留原始 ASR 字幕，再输出翻译语言字幕；翻译失败时不会删除原始字幕。
• 识别和翻译进度增加预估剩余时间，运行日志默认展示关键进度和问题提示，详细日志用于排查底层细节。
• 界面语言扩展为简体中文、繁体中文、英语、日语、韩语和西班牙语；首次启动默认跟随系统语言，手动切换后会记住。
• 优化 CUDA/GPU、模型下载、FFmpeg、权限和磁盘空间等常见问题的用户提示。
• 更新中英文 README、Wiki 文档和预览截图，明确本地 AI 字幕生成、SRT 翻译、GUI/CLI 和隐私边界。

发布资产

• AiSRT-v0.1.2-windows-portable.zip：轻量 Windows 包，包含源码、启动脚本和依赖文件。
• aisrt-0.1.2-py3-none-any.whl：Python wheel，适合手动安装的用户。
• SHA256SUMS.txt：发布资产 SHA256 校验值。

注意事项

• 发布包不包含 Python、PyTorch/CUDA 运行库、模型权重、模型缓存、FFmpeg、媒体文件、生成字幕、截图、测试文件或日志。
• 解压 ZIP 后运行 install_runtime.bat，脚本会创建 .venv 并安装 Python 依赖。
• 使用远程模型 ID 时，AISRT 会在首次运行时下载模型权重到配置的 Hugging Face 缓存目录。
• FFmpeg 和 ffprobe 需要单独安装，并确保可在 PATH 中直接运行。

English

This release provides a lightweight Windows portable source package and a Python wheel, with improvements focused on local SRT translation, GUI multilingual UX, and clearer logs.

Highlights

• Added local SRT translation for both the GUI and the ai-sub-translate CLI.
• The main GUI flow can now recognize and translate in one run, keeping the original ASR subtitles before writing translated subtitles; translation failures no longer remove the original subtitles.
• Recognition and translation progress now include estimated remaining time. The default run log shows key progress and issue hints, while detailed logs remain available for troubleshooting.
• UI copy now supports Simplified Chinese, Traditional Chinese, English, Japanese, Korean, and Spanish. First launch follows the system language when supported, and manual changes are remembered.
• Improved user-facing hints for CUDA/GPU, model download, FFmpeg, permission, and disk-space problems.
• Refreshed the Chinese and English README files, Wiki docs, and preview screenshots around local AI subtitle generation, SRT translation, GUI/CLI usage, and privacy boundaries.

Assets

• AiSRT-v0.1.2-windows-portable.zip: lightweight Windows package with source code, launch scripts and dependency files.
• aisrt-0.1.2-py3-none-any.whl: Python wheel for users who prefer manual Python installation.
• SHA256SUMS.txt: SHA256 checksums for release assets.

Notes

• Python, PyTorch/CUDA runtime libraries, model weights, model caches, FFmpeg, media files, generated subtitles, screenshots, tests and logs are not included.
• Run install_runtime.bat after extracting the ZIP to create .venv and install Python dependencies.
• When remote model IDs are used, AISRT downloads model weights to the configured Hugging Face cache on first use.
• FFmpeg and ffprobe must be installed separately and available in PATH.

AISRT v0.1.1

AISRT v0.1.1

简体中文

本次发布提供轻量 Windows portable 源码包和 Python wheel。

主要变化

• 补齐 GitHub 开源协作文件：Issue 模板、PR 模板、行为准则、支持说明、轻量 CI 和 Dependabot 配置。
• 优化 Python 环境配置，将默认 CUDA PyTorch 固定项拆分到 requirements-torch-cu130.txt，运行环境和开发环境复用同一套配置。
• 增强开源卫生检查，覆盖社区文件、requirements 文件、Markdown 相对链接和常见隐私信息模式。
• 修复 Windows portable 打包脚本，避免 Windows PowerShell 兼容性问题和空 ZIP 包误发布。
• 移除 README 中指向本地截图目录的引用，因为截图属于发布前应忽略的运行产物。

发布资产

• AiSRT-v0.1.1-windows-portable.zip：轻量 Windows 包，包含源码、启动脚本和依赖文件。
• aisrt-0.1.1-py3-none-any.whl：Python wheel，适合手动安装的用户。
• SHA256SUMS.txt：发布资产 SHA256 校验值。

注意事项

• 发布包不包含 Python、PyTorch/CUDA 运行库、模型权重、模型缓存、FFmpeg、媒体文件、生成字幕、截图、测试文件或日志。
• 解压 ZIP 后运行 install_runtime.bat，脚本会创建 .venv 并安装 Python 依赖。
• 使用远程模型 ID 时，AISRT 会在首次运行时下载模型权重到配置的 Hugging Face 缓存目录。
• FFmpeg 和 ffprobe 需要单独安装，并确保可在 PATH 中直接运行。

English

This release provides a lightweight Windows portable source package and a Python wheel.

Highlights

• Added GitHub open-source collaboration files: issue templates, pull request template, code of conduct, support guide, lightweight CI and Dependabot configuration.
• Improved Python environment setup by moving the default CUDA PyTorch pins into requirements-torch-cu130.txt, shared by runtime and development installs.
• Strengthened open-source hygiene checks for community files, requirements files, Markdown relative links and common private data patterns.
• Fixed the Windows portable packaging script to avoid Windows PowerShell compatibility issues and prevent empty ZIP releases.
• Removed the README reference to local screenshots because screenshots are runtime artifacts ignored before release.

Assets

• AiSRT-v0.1.1-windows-portable.zip: lightweight Windows package with source code, launch scripts and dependency files.
• aisrt-0.1.1-py3-none-any.whl: Python wheel for users who prefer manual Python installation.
• SHA256SUMS.txt: SHA256 checksums for release assets.

Notes

• Python, PyTorch/CUDA runtime libraries, model weights, model caches, FFmpeg, media files, generated subtitles, screenshots, tests and logs are not included.
• Run install_runtime.bat after extracting the ZIP to create .venv and install Python dependencies.
• When remote model IDs are used, AISRT downloads model weights to the configured Hugging Face cache on first use.
• FFmpeg and ffprobe must be installed separately and available in PATH.

AISRT v0.1.0

AISRT v0.1.0

This release provides a lightweight Windows portable source package and a Python wheel.

Assets:

• AiSRT-v0.1.0-windows-portable.zip: lightweight Windows package with source code, launch scripts and dependency files.
• aisrt-0.1.0-py3-none-any.whl: Python wheel for users who prefer manual Python installation.
• SHA256SUMS.txt: SHA256 checksums for release assets.

Notes:

• Python, PyTorch/CUDA runtime libraries, model weights, model caches, FFmpeg, media files, generated subtitles, screenshots, tests and logs are not included.
• Run install_runtime.bat after extracting the ZIP to create .venv and install Python dependencies.
• When remote model IDs are used, AISRT downloads model weights to the configured Hugging Face cache on first use.
• FFmpeg and ffprobe must be installed separately and available in PATH.