DATA-AI 是 AI Skill Hub 本期精选Agent工作流之一。综合评分 8.0 分,整体质量较高。我们强烈推荐将其纳入你的 AI 工具库,帮助提升工作效率。
DATA-AI 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。
DATA-AI 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。
# 方式一:pip 安装(推荐)
pip install data-ai
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install data-ai
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/badhope/DATA-AI
cd DATA-AI
pip install -e .
# 验证安装
python -c "import data_ai; print('安装成功')"
# 命令行使用
data-ai --help
# 基本用法
data-ai input_file -o output_file
# Python 代码中调用
import data_ai
# 示例
result = data_ai.process("input")
print(result)
# data-ai 配置文件示例(config.yml) app: name: "data-ai" debug: false log_level: "INFO" # 运行时指定配置文件 data-ai --config config.yml # 或通过环境变量配置 export DATA_AI_API_KEY="your-key" export DATA_AI_OUTPUT_DIR="./output"
🌐 English | 中文
| Category | Features |
|----------|----------|
| Chat | SSE streaming, Markdown code blocks, image attachments, reasoning stream, virtualised long sessions |
| Tools | Read/write, apply_edit with tree-sitter syntax gate, grep/glob, shell, background jobs, Jupyter, web search/fetch, Git suite, worktrees |
| Planning | Multi-step plan cards, parallel sub-agents (up to 4), declarative .claude/agents/*.md, live execution graph |
| Approval | Six modes (incl. plan-only / approve-plan), wildcard rule engine, three-way approval card with diff preview |
| Permissions | Four sandbox levels, protected paths, credential-path guard, token-level shell checks, MCP least-privilege allowlists |
| Memory | AGENTS.md (global ~/.agents + project + nested), vector long-term memory, auto-extraction, web management |
| Knowledge | Import docs, BM25 retrieval injected per turn as read-only reference — no extra vector store |
| Persistence | Sessions, checkpoint rollback (works without git), hot config reload, token usage stats, full-text session search |
| Models | Any OpenAI-compatible endpoint, Ollama, multi-profile switching, small_model for cheap auxiliary calls, --probe endpoint audit |
| Extension | MCP client (stdio/SSE/HTTP + OAuth), MCP registry with one-click add, skills .spark/skills/, slash commands, plugins, hooks |
| Interface | zh/en, light/dark, command palette, rebindable shortcuts, file viewer, terminal, web reader, execution graph, desktop notifications |
| Workflow | spark review (AI diff audit), spark exec --json, !cmd inline shell, single-file HTML session sharing, session import, spark update |
| Operations | doctor --fix self-healing, doctor --backup (VACUUM INTO), structured audit log, CSP + security headers, auto-update notify |
Any OpenAI-compatible endpoint (DeepSeek / Qwen / GLM / Kimi / Doubao / Ollama /
custom), saved model profiles with one-click switching, a Test connection probe, and
three reasoning-effort levels mapped to provider-native parameters.

Four access levels with explicit behaviour descriptions. The protected-path list is
user-editable; .git/hooks, ~/.spark and Spark's own source are always protected.

DATA-AI是一个功能强大的开源AI工作流框架,支持高级数据分析和可视化
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。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
经综合评估,DATA-AI 在Agent工作流赛道中表现稳健,质量优秀。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 | DATA-AI |
| 原始描述 | 开源AI工作流:A comprehensive AI assistant framework built on LangChain ecosystem with advance。⭐15 · Python |
| Topics | ai-agentautomationdata-analysisdata-visualizationfastapi |
| GitHub | https://github.com/badhope/DATA-AI |
| License | MIT |
| 语言 | Python |
收录时间:2026-05-26 · 更新时间:2026-05-30 · License:MIT · 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