v2.5.0

🚀 Nexent：开源智能体平台 / Nexent: Open Source Intelligent Agent Platform

我们很高兴地宣布Nexent v2.5.0 正式发布！🎉

Nexent 是一个开源智能体平台，能够将流程的自然语言转化为完整的多模态智能体 —— 无需编排，无需复杂拖拉拽。基于 MCP 工具生态，Nexent 提供强大的模型集成、数据处理、知识库管理、零代码智能体开发能力。我们的目标很简单：将数据、模型和工具整合到一个智能中心中，使日常工作流程更智能、更互联。

We are excited to announce that Nexent v2.5.0 is released! 🎉

Nexent is an open-source agent platform that turns process-level natural language into complete multimodal agents — no diagrams, no wiring. Built on the MCP tool ecosystem, Nexent provides model integration, data processing, knowledge-base management, and zero-code agent development. Our goal is simple: to bring data, models, and tools together in one smart hub, making daily workflows smarter and more connected.

---

新功能 / New Features

• 重构 Agent 配置与 NL2Agent 生成流程，支持从自然语言需求创建 Agent、发现并安装可用资源，以及推荐和预填已绑定工具。 / Refactored Agent configuration and NL2Agent generation to create agents from natural-language requirements, discover and install available resources, and recommend and prefill bound tools.
• 新增完整的 Agent 评估工作台，覆盖评估集、评估器、标注、调度执行与中英文报告，并提供通用基准测试和上下文对比工具。 / Added an end-to-end agent evaluation workspace covering datasets, evaluators, annotations, scheduled execution, and bilingual reports, together with generic benchmarking and context-comparison tools.
• 新增长期记忆 Dreaming 与版本化管理能力，并提供记忆配置、检索记录和可视化管理界面。 / Added long-term memory Dreaming and version management, with configuration, retrieval records, and visual management interfaces.
• 扩展运行时沙箱，支持会话级与系统级隔离、默认启用系统级沙箱、在沙箱内执行 Skill 脚本，并在前端展示生成文件。 / Expanded runtime sandboxing with session-level and system-level isolation, system sandboxing enabled by default, Skill script execution inside sandboxes, and generated-file display in the frontend.
• 新增 S3 文件上传下载工具并增强文件处理链路，使 Skill 和工具能够直接处理本地文件与产物。 / Added S3 upload and download tools and enhanced the file-processing pipeline so Skills and tools can work directly with local files and generated artifacts.
• 扩展统一模型网关，接入 LLM、Embedding、Rerank 及图像、视频、音频理解能力，并增强 ASR、结构化 VLM 消息、Logprobs 与瞬态错误重试支持。 / Expanded the unified model gateway across LLM, embedding, reranking, and image, video, and audio understanding, with improved ASR, structured VLM messages, logprobs, and transient-error retries.
• 增强知识库与资源治理，支持个人知识库容量、租户资源限额、会话级知识范围选择及更灵活的文件切分与处理。 / Enhanced knowledge-base and resource governance with personal capacity controls, tenant quotas, conversation-level knowledge scoping, and more flexible file splitting and processing.
• 优化对话体验，支持消息时间戳、按需分页加载、批量删除会话，并增强新旧聊天界面的运行时协同。 / Improved conversations with message timestamps, on-demand pagination, bulk deletion, and stronger runtime coordination across the new and legacy chat interfaces.
• 扩展平台开放能力，新增 API Key 管理、北向 /models 与 /generate_title 接口，并补充运行接口的会话创建事件。 / Expanded platform APIs with API key management, northbound /models and /generate_title endpoints, and conversation-creation events in the run API.
• 新增上下文与缓存监控指标，并强化元数据传递和提示注入防护，提升 Agent 运行的可观测性与安全性。 / Added context and cache metrics while strengthening metadata propagation and prompt-injection protection for more observable and secure agent execution.

Bug 修复 / Bug Fixes

• 修复知识库导入死锁、不可重试错误未及时终止、文件删除与预览异常、EPUB 处理失败及数值检索排序问题。 / Fixed knowledge-base import deadlocks, non-retryable failures that did not terminate promptly, file deletion and preview errors, EPUB processing failures, and numeric retrieval ranking.
• 修复 NL2Agent 表单锁定与工具配置问题，并增强 Skill 编码、中文名称、重命名元数据及 Agent 导入复制时的冲突处理。 / Fixed NL2Agent form locking and tool configuration, and improved Skill encoding, Chinese names, rename metadata, and conflict handling during agent import and copy.
• 修复会话删除、重命名、恢复及分页后空列表问题，并解决流式响应缓存、空最终回答和运行时内存压力异常。 / Fixed conversation deletion, renaming, resume, and empty lists after pagination, along with streaming cache issues, empty final answers, and runtime memory pressure.
• 修复评估任务在沙箱中的执行、中文评估报告、场景描述校验及评估路由和权限配置问题。 / Fixed evaluation execution in sandboxes, Chinese evaluation reports, scenario-description validation, and evaluation routing and permission configuration.
• 修复 MCP 状态显示、工具刷新、Registry 删除及配置页同步问题，同时完善容器启动日志展示。 / Fixed MCP status display, tool refresh, Registry deletion, and configuration-page synchronization, while improving container startup log visibility.
• 修复定时任务无法创建、模型管理会话过期处理以及外部 A2A 子 Agent 凭据传递等运行与鉴权问题。 / Fixed scheduled-task creation, expired-session handling in model management, and credential propagation to external A2A sub-agents.

文档与部署变化 / Docs & Deployment Changes

• 更新 Docker、Helm、离线包与镜像构建配置，改进 MinIO 数据持久化和就绪检查，并保持环境变量配置兼容性。 / Updated Docker, Helm, offline-package, and image-build configurations, improved MinIO persistence and readiness checks, and preserved environment-variable compatibility.
• 汇总 v2.5.0 SQL 迁移并增强合并迁移的重复应用逻辑，同时补充不可变迁移文件规范。 / Consolidated v2.5.0 SQL migrations, improved reapplication of changed merged migrations, and documented the immutable-migration policy.
• 更新安装、Kubernetes、主页、资源管理和官方文件生成 Skill 文档，并补充账户初始化与自定义文件生成指南。 / Updated installation, Kubernetes, home-page, resource-management, and official file-generation Skill documentation, with new account-setup and custom file-generation guides.

What's Changed

• feat(observability): expose context and cache metrics by @wuyuanfr in #3572
• Implement named offline package uploads to Huawei OBS by @hhhhsc701 in #3608
• Bugfix: Fix inability to rename conversation name by @xuyaqist in #3613
• fix: provide local import allowlist before code execution by @EDDIWARD in #3614
• Bugfix: Resolve streaming response caching problem by @xuyaqist in #3623
• Bugfix: Fix MCP tool status display showing "enabled" when enabled is false by @xuyaqist in #3632
• 🐛 fix : restore ASSET_OWNER nav permissions and improve group selection UX by @Lifeng-Chen in #3629
• ♻️ Refactor: NL2Skills using new thread UI by @Jasonxia007 in #3631
• Update doc: resource management by @xuyaqist in #3607
• Bugfix: Fix inability to delete conversation by @xuyaqist in #3635
• Bugfix: normalize repository tags as arrays by @Summer-Si in #3648
• bugfix: 中英文词条长度限制调整，接口权限判断逻辑调整，支持jpg格式图片 by @hzwllm in #3628
• bugfix:系统中显示的时间对齐用户本地时间 by @hzwllm in #3647
• feat: account MinIO in knowledge-base quotas by @MoeexT in #3642
• ✨ Feature(agent-evaluation): add UT suite, wire routers & scheduler, decla… by @cj2026-bit in #3622
• 🐛 Bugfix: Fixed an issue where the tenant resource management page failed to correctly detect model connectivity. by @YehongPan in #3651
• 🐛 Bugfix: Fixed an issue where clicking to view skills before the skills API returned a response would result in a blank screen on the agent editing page. by @YehongPan in #3663
• 🐛 Bugfix: Fixed an issue where an embedding model could not be added if the URL did not end with /embeddings. by @YehongPan in #3655
• feat(memory): Dreaming and versioned long-term memory by @JasonW404 in #3509
• fix: resolve embedding_model_id from knowledge record in data process… by @Lifeng-Chen in #3653
• fix: pass security credentials to external A2A sub-agents by @2862282695gjh-afk in #3636
• Bugfix: Fix markdown table rendering issue after reasoning completes by @xuyaqist in #3643
• Configure CAS settings for renewal interval and session handling by @hhhhsc701 in #3666
• test(memory): harden evidence schema compatibility harness by @JasonW404 in #3669
• chore: remove spec-coding skill by @JasonW404 in #3672
• ✨ Parameterize file split function in data-process task by @Jasonxia007 in #3665
• 🐛 Bugfix: fix status tag icon misalignment and localize auth dialog slogan by @ayiya12 in #3673
• enforce tenant resource limits by @Junqilee in #3670
• feat(gateway): wire VLM image/video/audio understanding via adapter by @jeffwu-1999 in #3668
• Bugfix: export agent with aidp skill failed by @bernard1234 in #3638
• test: mock tenant resource limit error by @Junqilee in #3686
• ✨feat:Enhance conversation knowledge base selection and rendering by @DongJiBao2001 in #3687
• feat(benchmark): add generic evaluation and comparison harness by @wuyuanfr in #3640
• 🐛 Bugfix: prevent empty agent final answers by @YehongPan in #3688
• docs: update home page documentation and add account setup guide by @charmingchi1 in #3662
• fix: correct Dreaming permission migration by @JasonW404 in #3693
• 🐛 Bugfix: Fix plan params error by @WMC001 in #3703
• Add new committer by @WMC001 in #3704
• fix(evaluation): align evaluation route and permission SQL by @cj2026-bit in #3683
• fix: improve numeric knowledge retrieval ranking by @EDDIWARD in #3621
• fix(deploy): eliminate redundant MinIO anonymous volumes by @Dallas98 in #3714
• test: make dreaming age fixture time-stable by @JasonW404 in #3715
• 修复部分依赖在源中找不到 by @hhhhsc701 in #3720
• feat(gateway): wire text/retrieval (llm + embedding + rerank) via adapter by @jeffwu-1999 in #3695
• fix(modelengine): preserve structured VLM messages by @JasonW404 in #3716
• ✨ Feature: Added tools for uploading and downloading files from S3 buckets, ensuring that skills and tools can directly process local files. by @YehongPan in #3724
• fix: remove erroneously added context_policy column from init.sql by @JasonW404 in #3517
• 🐛 Bugfix: remove delete account button from avatar dropdown by @ayiya12 in #3727
• 🐛 Bugfix: remove obsolete skill repository uniqueness constraint by @Summer-Si in #3641
• 智能体对话转发到runtime by @hhhhsc701 in #3721
• ✨Feature: Agent配置重构&NL2Agent by @xuyaqist in #3726
• 🐛 Bugfix: Skill encoding, skill file read error, skills with chinese name by @Jasonxia007 in #3718
• 将每次chunk存一次数据库改为最终统一存数据库，redis不变 by @hhhhsc701 in #3733
• Bugfix: resolve skill conflicts during agent import by @Summer-Si in #3734
• ✨ feature personal kb capacity by @cj2026-bit in #3712
• fix: sync MCP add/edit pages with config page, display startup logs, and resolve tool refresh bug by @menglinghan in #3735
• ♻️ Add re-apply logic when merged sql files were changed by @WMC001 in #3737
• fix: keep greeting while exporting & importing agents by @MoeexT in #3736
• 🐛 Bugfix: change color of PREVIEW button to blue by @WMC001 in #3741
• 🐛 Bugfix: can not preview json file by @WMC001 in #3742
• ✨feat:Add metadata parameter mechanism and fix prompt injection issues by @DongJiBao2001 in #3738
• fix(frontend): reorder displayed citation indexes by @JasonW404 in #3739
• Bugfix: Agent配置页面问题修复 by @xuyaqist in #3744
• ✨ Support api-key management by @Jasonxia007 in #3745
• feat(modelengine): adapt ASR audio models to vlm4 slot via modelengine provider by @jeffwu-1999 in #3743
• Document immutable SQL migration file policy by @hhhhsc701 in #3753
• ✨ Feature: Enhanced file processing logic by @YehongPan in #3746
• Dc/fix/tool default value ranges 20260819：fix(tool config): 统一工具参数边界约束校验（SDK + Backend + Frontend） by @charmingchi1 in #3729
• bigfix: add DashScope VLM adapter for audio via vlm slot by @jeffwu-1999 in #3755
• 🐛 Bugfix: Failed to process .epub because of missing pandoc by @Jasonxia007 in #3760
• Feat: 新增引导指引用户使用nl2agent功能生成Agent by @xuyaqist in #3756
• Bugfix: Preserve Skill frontmatter when renaming during Agent import by @Summer-Si in #3763
• 🐛 Bugfix: Fix NL2Agent tool configuration and form unlock flow by @RedShakespeare in #3761
• ✨ Feature: Supports session-level sandbox. by @YehongPan in #3764
• Bugfix🐛为对话列表和知识库列表增加按需分页加载机制 by @DoYX in #3719
• Quote Docker environment startup mode by @hhhhsc701 in #3765
• feat(models): add exponential backoff retry for transient LLM errors by @TiTi-iTiT in #3747
• 🐛 Fix: Failed Knowledge-Base File Deletion, Error Details, and Preview Styling by @cj2026-bit in #3750
• ✨feat:Adapt AIDP to new agent configuration page and fix bugs by @DongJiBao2001 in #3762
• 🐛Bugfix: improve resource management filtering and iData search support by @ayiya12 in #3768
• fix(auth): handle session expiry with 401 responses across model management by @jeffwu-1999 in #3766
• ✨ Bugfix: fix chat resume by @xuyaqist in #3770
• fix: update interactive creation assistant title by @Summer-Si in #3769
• bugfix: support skill conflict resolution on agent copy by @bernard1234 in #3767
• ✨ Add northbound api: /models /generate_title by @Jasonxia007 in #3771
• fix(agentConfigService): 为代理配置参数添加约束校验支持 by @charmingchi1 in #3773
• ✨ Feature: Add NL2Agent installable resource discovery and installation flow by @RedShakespeare in #3775
• ✨ Feature:Frontend display for generated files and sandbox optimization by @YehongPan in #3774
• fix: delete MCP Registry by @menglinghan in #3757
• Upgrade to version v2.5.0 by @hhhhsc701 in #3776
• fix(frontend): 补齐用户组多选框搜索能力并适配 antd6 showSearch 配置 by @charmingchi1 in #3777
• Feat: Show hint that default editing is in draft mode by @xuyaqist in #3772
• ✨Feature: 在新旧版对话窗口中添加消息时间戳 by @DoYX in #3732
• fix: Unable to create scheduled tasks by @Dallas98 in #3780
• ✨ Add logprobs param to llm if its config has been added by @WMC001 in #3779
• ✨ Feature: System-level sandboxing is enabled by default. by @YehongPan in #3782
• 🐛 Fix: fail knowledge ingestion on non-retryable errors by @cj2026-bit in #3778
• 🐛Fix: localize resource deletion confirmations by @ayiya12 in #3781
• Bugfix: 修复conversation无法删除的问题，设置修改conversation title的样式 by @xuyaqist in #3785
• 🐛 Bugfix: Unlock NL2Agent form after prompt updates and stops by @RedShakespeare in #3788
• ✨ Feature: Recommend and prefill bound NL2Agent tools by @RedShakespeare in #3783
• Refactor: 重构chat页面分页加载功能，默认加载30条，修复删除对话后列表为空的问题 by @xuyaqist in #3790
• 🐛 Fix: kb import deadlock problem by @cj2026-bit in #3786
• 🐛 Fix: support Chinese evaluation reports and validate AI evaluation scene descriptions by @cj2026-bit in #3791
• ✨ Enhance read_skill_md tool with encoding utf-8 & gb18030 by @Jasonxia007 in #3784
• 🐛 Bugfix: Add manual NL2Agent unlock action by @RedShakespeare in #3800
• feat(chat): 会话列表支持批量删除 by @charmingchi1 in #3792
• fix(runtime): bound streaming memory pressure by @JasonW404 in #3801
• ✨ Feature: Supports the execution of SKILL scripts within a sandbox. by @YehongPan in #3802
• 🐛 Fix: agent evaluation execution in the runtime sandbox by @cj2026-bit in #3799
• ✨ Improve memory Dreaming configuration UI by @JasonW404 in #3804
• Bugfix: 支持在北向接口的run接口中返回conversation_created的chunk by @xuyaqist in #3793

New Contributors

• @EDDIWARD made their first contribution in #3614
• @2862282695gjh-afk made their first contribution in #3636
• @Junqilee made their first contribution in #3670
• @charmingchi1 made their first contribution in #3662
• @DoYX made their first contribution in #3719
• @TiTi-iTiT made their first contribution in #3747

Full Changelog: v2.4.1...v2.5.0