Repository navigation
[4.30.0-beta.1] - 2026-11-18
本版本为 Beta 测试版,可能存在不稳定或兼容性问题,建议先在测试环境验证。本次涉及对话数据库、Agent 配置和会话设置迁移。升级前请备份数据与配置;回退旧版需同时恢复升级前备份。
自 v4.30.0-alpha.1 以来的更新
- 用户追加指令(Steering)时,自动将等待中的工具调用移至后台任务,前台继续处理新输入。 (#11142)
- 修复 QQ 官方群聊流式回复被合并的问题。 (#11229)
- QQ 官方适配器支持获取私聊用户名字,群聊显示名称优先使用群内用户昵称。 (#11184)
- 插件 Views 能力升级:ChatUI 新增分页工作台与常驻插件视图,支持从「+」打开视图、拖动调整宽度、全屏以及按会话恢复标签页和布局。 (#11218)
- 插件 Views 能力升级:插件可在聊天消息内展示交互式 HTML 组件,通过命令、消息处理器或模型工具创建,并处理用户操作;其他平台保留文本回退。 (#11179)
- 插件 Views 能力升级:插件视图统一在
metadata.yaml的views中声明。 (#11218) - ChatUI 支持语音转写为可编辑草稿;ChatUI 内模型配置页上方增加「对话 / 语音转文字 / 文字转语音」分类入口。 (#11170)
- 将聊天推理侧栏改为消息内可折叠的运行详情,分组展示思考、工具调用与中间回复;后端持续运行时,刷新或重连可恢复正在执行的对话并保留后续回复。 (#11166)
- 支持 Gemini 4 系列的推理强度配置,完善预览模型和请求参数兼容性。 (#11219)
- 减少上下文 Token 估算与截断的重复计算,合并会话元数据与幂等性查询,优化性能。 (#11167, #11199)
- 定时任务和后台任务唤醒时,仅通过显式消息工具发送通知,避免额外自动发送完成摘要。 (#11198)
- 保留 OpenRouter 非流式回复中的推理详情。 (#11168)
- 修复 Gemini 历史消息中的 URL 音频处理,并保留 SDK 3.4 下的自定义 User-Agent。 (#11143, #11196)
- DashScope TTS 支持 Qwen-Audio-TTS。 (#10495)
- System One 布尔判断支持可选的真假判定条件,修复部分兼容端点连接测试返回 HTTP 400 的问题。 (#11062)
- 修复切换聊天会话时欢迎页闪现的问题,并移除已弃用的 VoceChat 图标引用。 (#11182, #11186)
- 测试配置与本地 AstrBot 数据隔离,完善 Windows 数据库清理,并在项目交付测试中显式禁用 Git 签名。 (#11231, #11222)
- 更新插件开发文档中的 QQ 群信息。 (#11178)
alpha.1 主要更新
- 将会话后端迁移到事件溯源方式,减少重复上下文写入和部分写放大,优化存储性能;增量保存消息与工具调用,完善执行中断后的历史处理。 (#10077, #11158, #11162)
- 引入新的 Agent 运行时模型。插件可通过
ctx.agent.run_turn()唤起 Agent turn。 (#10934, #10936, #11052, #11148, #11153) - 重构子代理为具有独立上下文的后台任务,支持通用与专用 Agent、追加指令和结果回传。入口由侧边栏移至 配置文件 → AI → 能力 → 子代理,按配置文件管理。 (#11053)
- 将自定义规则与会话管理统一至 配置文件 → 管理会话配置映射,支持基础配置与字段覆盖、批量编辑和恢复继承;模型指令与页面共用同一份设置。 (#11097)
- 新增 Windows AppContainer 本机沙盒,支持工作区隔离、代码执行与联网控制。 (#10466)
- 新增 DeepSeek 语音识别与语音合成提供商。 (#11118)
- 新增 OpenAI Decisions 判断模型,支持选择、布尔判断与评分。 (#10968)
- 欢迎页 → 总览:展示每日用量、系统资源及机器人异常。 (#11100)
- 支持在外观设置中将聊天入口放入侧边栏。 (#11090)
- 升级 Markdown 渲染,按需加载聊天与语法高亮资源。 (#11138)
- 插件配置支持可折叠分组,并可声明默认收起状态。 (#11067)
- 插件页面可获取当前主题颜色及其变化。 (#11046)
- 重整 AI 开发文档,补充插件配置控件示例与截图。 (#11154)
alpha.1 优化与修复(完整保留)
- 完善运行中追加输入的保存与继续执行。 (#11141)
- 跳过运行时不可用的子代理配置,避免阻断主会话。 (#11128)
- 允许取消选择已禁用、缺失或运行时不可用的子代理配置。 (#11147)
- 移除完整工具 Schema 模式下的默认工具调用提示词。 (#11152)
- 修复模型最后一次重试成功仍被判定为失败的问题。 (#10922)
- 修复 Anthropic 非流式响应丢失多个文本块的问题。 (#11057)
- 搜索密钥失效或额度不足时,尝试其他已配置的密钥。 (#10877)
- 将 Gemini TTS 默认模型更新为
gemini-3.8-flash-tts。 (#10828) - 修复 QQ 按钮回复丢失图片、文件等媒体的问题。 (#10903)
- 避免 QQ 图片上传失败时记录完整图片载荷,减少日志对面板的影响。 (#10740)
- 优化 OneBot 流式分段,减少长文本的重复扫描。 (#10989)
- Telegram 发送者没有用户名时,使用显示名称。 (#10715)
- 修复 Firefox 中聊天历史滚动跳动的问题。 (#11036)
- 修复编辑模型配置时预览与能力标识不一致的问题。 (#10792)
- 统一 WebUI 菜单、弹窗样式,优化滚动条和点击反馈。 (#11107, #11090, #11102)
- 修复 EPUB 解析丢失链接文字的问题。 (#10794)
- 保留 GIF 转换为拼图后的说明,避免历史上下文丢失动画信息。 (#10837)
- 将插件加载及执行失败日志归属到对应插件,方便筛选排查。 (#10904)
- SSE 流关闭时通知插件页面,避免连接状态滞留。 (#10932)
- 控制台日志时间戳增加日期。 (#9617)
This is a Beta release and may have stability or compatibility issues. Validate it in a test environment first. This release migrates conversation storage, Agent configurations, and session settings. Back up data and configuration before upgrading; downgrading also requires restoring the pre-upgrade backup.
Changes since v4.30.0-alpha.1
- Add a tabbed ChatUI workbench and resident plugin views, with a + launcher, resizable panels, fullscreen mode, and per-session restoration of tabs and layout. (#11218)
- Let plugins render interactive HTML components inside chat messages, create them through commands, message handlers, or model tools, and handle user actions. Preserve text fallback on other platforms. (#11179)
- Declare plugin views in
metadata.yamlunderviews. (#11218) - Merge Plugin Views and ChatUI Cards documentation into a Plugin Views section organized around standalone pages, message components, and workbenches. Add bilingual tutorials, screenshots, and examples, with redirects from previous links. (#11179, #11218)
- Add microphone dictation into editable drafts and direct recording and sending of audio messages. Replace the unified model configuration selector with Conversation / Speech to Text / Text to Speech categories. (#11170)
- Replace the reasoning sidebar with collapsible inline run details grouping thinking, tool calls, and intermediate replies. Resume active conversations after refresh or reconnect and preserve subsequent replies while the backend remains running. (#11166)
- Move supported pending native tools to the background when follow-up input arrives so the foreground can process it. Preserve caller identity, permissions, and complete background results, wake the original conversation on completion, and cancel transferred work when the session stops. (#11142)
- Support QQ Official sender profiles, avatars, and on-demand user information queries, preferring group nicknames for group display names. (#11184)
- Support Gemini 4 reasoning-effort configuration and improve preview-model and request-parameter compatibility. (#11219)
- Reduce repeated context token estimation and truncation work, and merge conversation metadata and idempotency lookups. (#11167, #11199)
- Send scheduled and background-task notifications through explicit message tools, avoiding extra automatic completion summaries. (#11198)
- Prevent QQ Official group streaming replies from being merged. (#11229)
- Preserve reasoning details in non-streaming OpenRouter responses. (#11168)
- Fix URL audio handling in Gemini history and preserve the custom User-Agent with SDK 3.4. (#11143, #11196)
- Support Qwen-Audio-TTS in the DashScope TTS provider. (#10495)
- Support optional true/false criteria for System One boolean decisions and fix HTTP 400 connection-test failures on compatible endpoints that require them. (#11062)
- Prevent the welcome screen from flashing when switching chat sessions and remove the legacy VoceChat logo reference. (#11182, #11186)
- Isolate test configuration from local AstrBot data, improve Windows database cleanup, and explicitly disable Git signing in project delivery tests. (#11231, #11222)
- Update QQ group information in the plugin development documentation. (#11178)
alpha.1 Highlights
- Migrate conversations to event sourcing, reducing duplicate context writes and some write amplification. Persist messages and tool calls incrementally and improve history handling after interrupted execution. (#10077, #11158, #11162)
- Introduce a new Agent runtime model. Plugins can invoke an Agent turn through
ctx.agent.run_turn(). (#10934, #10936, #11052, #11148, #11153) - Rebuild subagents as background tasks with independent context, generic or specialized configurations, follow-up inputs, and result delivery. Move management from the sidebar to Config → AI → Capabilities → Subagents, with per-profile settings. (#11053)
- Merge Custom Rules and Session Management into Config → Manage session mappings, supporting base profiles, field overrides, batch editing, and restored inheritance. Model commands and the editor share the same settings. (#11097)
- Add a Windows AppContainer sandbox with workspace isolation, code execution, and network controls. (#10466)
- Add DeepSeek speech recognition and synthesis providers. (#11118)
- Add OpenAI Decisions models for choices, boolean questions, and scores. (#10968)
- Welcome → Overview now shows daily usage, system resources, and bot errors. (#11100)
- Add an Appearance setting to show Chat in the sidebar. (#11090)
- Upgrade Markdown rendering and load chat and syntax-highlighting assets on demand. (#11138)
- Add collapsible plugin configuration groups with configurable initial state. (#11067)
- Expose current theme colors and their updates to plugin pages. (#11046)
- Reorganize AI development guides and add plugin configuration control examples and screenshots. (#11154)
alpha.1 Improvements and Fixes (retained in full)
- Improve persistence and continuation of inputs added during execution. (#11141)
- Skip subagent configurations with unavailable runtimes so they do not block the main conversation. (#11128)
- Allow deselecting disabled, missing, or runtime-unavailable subagent configurations. (#11147)
- Remove default tool-call instructions in full-schema mode. (#11152)
- Fix successful final model retries being reported as failures. (#10922)
- Preserve all text blocks in non-streaming Anthropic responses. (#11057)
- Try other configured search keys when a key is invalid or its quota is exhausted. (#10877)
- Update the default Gemini TTS model to
gemini-3.8-flash-tts. (#10828) - Preserve images, files, and other media in QQ button replies. (#10903)
- Stop logging full image payloads on QQ upload failures, reducing their impact on the dashboard. (#10740)
- Improve OneBot streaming segmentation by avoiding repeated scans of long text. (#10989)
- Use Telegram display names when usernames are absent. (#10715)
- Fix Firefox chat history scroll jumps. (#11036)
- Preserve model previews and capability indicators while editing provider configurations. (#10792)
- Unify WebUI menus and dialogs and refine scrollbars and press feedback. (#11107, #11090, #11102)
- Preserve link text when parsing EPUB files. (#10794)
- Retain GIF montage guidance in conversation history so animation context is preserved. (#10837)
- Attribute plugin loading and execution errors to their plugins for easier log filtering. (#10904)
- Notify plugin pages when SSE streams close to avoid stale connection states. (#10932)
- Include dates in console log timestamps. (#9617)
Full Changelog: v4.29.4...v4.30.0-beta.1
Changes since alpha.1: v4.30.0-alpha.1...v4.30.0-beta.1
What's Changed
- fix: handle URL audio in Gemini history by @wcqqq1214 in #11143
- feat(chat): add inline run details and preserve replies after refresh by @Soulter in #11166
- feat(chat): add voice transcription and audio messages by @Soulter in #11170
- docs(plugin-new): change QQ Group by @mantoujun12 in #11178
- fix(chat): prevent welcome screen flash when switching sessions by @Soulter in #11182
- fix: support Qwen-Audio-TTS in DashScope provider by @C10H14N2O5 in #10495
- perf(context): reduce repeated work in token estimation and truncation by @whatevertogo in #11167
- fix: preserve Gemini user agent with SDK 3.4 by @wcqqq1214 in #11196
- fix: suppress automatic replies for cron and background wakeups by @RC-CHN in #11198
- perf(conversation): merge metadata and idempotency lookups by @RC-CHN in #11199
- feat(plugin): add interactive plugin message views by @Soulter in #11179
- feat(chatui): add tabbed workbench and resident plugin views by @Soulter in #11218
- fix(dashboard): Remove legacy VoceChat logo reference by @Sha1rholder in #11186
- feat: gemini 4 model series reasoning effort support without bumping google-genai by @Sha1rholder in #11219
- choe: remove project delivery test by @Sha1rholder in #11222
- feat: background pending tools when steering arrives by @Soulter in #11142
- fix: preserve OpenRouter reasoning details in non-streaming replies by @123XieC in #11168
- fix: prevent QQ official group streaming replies from being merged by @Soulter in #11229
- feat(qqofficial): support sender profiles and prefer group nicknames by @clown145 in #11184
- fix(provider): support optional System One boolean criteria by @mothieras in #11062
- fix: isolate test configuration from local AstrBot data by @Soulter in #11231
- chore: bump version to 4.30.0-beta.1 by @Soulter in #11233
New Contributors
- @Sha1rholder made their first contribution in #11186
- @123XieC made their first contribution in #11168
- @mothieras made their first contribution in #11062
Full Changelog: v4.30.0-alpha.1...v4.30.0-beta.1