开源AI工作流 是 AI Skill Hub 本期精选Agent工作流之一。综合评分 7.5 分,整体质量较高。我们推荐使用将其纳入你的 AI 工具库,帮助提升工作效率。
基于Shell的企业级AI工作流平台,专为SEA金融服务(印度尼西亚)设计,提供能力层企业代理平台
开源AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
基于Shell的企业级AI工作流平台,专为SEA金融服务(印度尼西亚)设计,提供能力层企业代理平台
开源AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 克隆仓库 git clone https://github.com/chaosxingxc-orion/spring-ai-ascend cd spring-ai-ascend # 查看安装说明 cat README.md # 按 README 完成环境依赖安装后即可使用
# 查看帮助 spring-ai-ascend --help # 基本运行 spring-ai-ascend [options] <input> # 详细使用说明请查阅文档 # https://github.com/chaosxingxc-orion/spring-ai-ascend
# spring-ai-ascend 配置说明 # 查看配置选项 spring-ai-ascend --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export SPRING_AI_ASCEND_CONFIG="/path/to/config.yml"
An open-source, enterprise-grade agent platform built for the Huawei Ascend (NPU) + Kunpeng (CPU) stack — on Spring AI, Spring Boot, and Java 21.
spring-ai-ascend lets a team stand up its own governed agent runtime the way it would stand up a Spring Boot service: import the BoM, override the SPI beans you care about, and ship. It is designed for self-hosting on Huawei silicon — Kunpeng (ARM64) for the JVM service tier and Ascend NPUs for model serving — so an enterprise can run the whole agent stack on its own hardware, OSS-first, with no proprietary-cloud lock-in.
What runs today vs. what's on the roadmap. The shipped runtime is a hardware-agnostic Spring AI / Java kernel — it runs on any JVM, and natively on Kunpeng/ARM64, so you develop and test anywhere. Ascend-NPU-optimised model serving and Kunpeng-tuned deployment profiles are the platform's design target, not yet shipped code. This boundary is marked honestly throughout; the machine-readable, per-capability ledger is docs/governance/architecture-status.yaml.
- Dual-mode orchestration. One runtime runs both deterministic graph state machines and ReAct-style agent loops, sharing a single interrupt primitive (SuspendSignal). A graph node can call an agent loop, which can call another graph — arbitrary bidirectional nesting, one Run lineage throughout. - Pluggable by SPI, not by patching. Memory, run persistence, model gateway, tool authorization, and resilience are SPI surfaces you implement and wire by dependency injection; in-memory reference implementations ship for local dev. - Multi-tenant + audit-grade. Every run carries a tenant id; storage-engine isolation, durable idempotency, and structured audit logging are first-class. - Posture-aware. dev is permissive for fast iteration; research/prod fail closed at startup when required configuration is missing.
./mvnw -T 1C -Pquality verify ```
Use verify, not test — test skips the *IT.java integration enforcers. -T 1C builds the reactor modules in parallel. Posture is selected by the APP_POSTURE environment variable (dev / research / prod); dev allows in-memory backends and only WARNs on missing config. The full boot-and-first-run walkthrough is in docs/quickstart.md.
```bash
该项目提供了一个基于Shell的企业级AI工作流平台,适合用于SEA金融服务的需求,值得关注
该工具未明确声明开源协议,商业使用前请联系原作者确认授权范围,避免侵权风险。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
经综合评估,开源AI工作流 在Agent工作流赛道中表现稳健,质量良好。如果你已有明确的使用需求,可以直接上手体验;如果还在评估阶段,建议对比同类工具后再做决策。
| 原始名称 | spring-ai-ascend |
| 原始描述 | 开源AI工作流:Capability-layer enterprise agent platform for SEA financial services (Indonesia。⭐6 · Shell |
| Topics | workflowshell |
| GitHub | https://github.com/chaosxingxc-orion/spring-ai-ascend |
| 语言 | Shell |
收录时间:2026-05-23 · 更新时间:2026-05-30 · License:未公布 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端