⚙️
Agent工作流

memgraph-ingester

基于 Java · 无代码搭建完整 AI 自动化流程
⭐ 6 Stars 💻 Java 📄 MIT 🏷 AI 7.1分
7.1AI 综合评分
workflowaiai-agentai-agentsclaude-codecodexjava
✦ AI Skill Hub 推荐

经 AI Skill Hub 精选评估,memgraph-ingester 获评「推荐使用」。这款Agent工作流在功能完整性、社区活跃度和易用性方面表现出色,AI 评分 7.1 分,适合有一定技术背景的用户使用。

📚 深度解析
memgraph-ingester 是一套完整的 AI Agent 自动化工作流方案。随着 AI 能力的不断提升,基于 Agent 的自动化工作流正在成为提升个人和团队效率的核心方式。区别于传统的 RPA 自动化(模拟鼠标键盘操作),AI Agent 工作流通过理解任务意图、动态规划执行路径,能够处理更复杂的非结构化任务。

memgraph-ingester 工作流的设计遵循"最小配置,最大复用"原则:核心逻辑已经封装好,用户只需配置自己的 API Key 和业务参数即可快速上手。工作流内置错误处理和重试机制,在网络波动或 API 限速等情况下仍能稳定运行,适合作为生产环境的自动化基础设施。

在实际部署时,建议先在测试环境中运行 3-5 次,验证各个环节的输出结果符合预期,再部署到生产环境。AI Skill Hub 评分 7.1 分,是同类 Agent 工作流中的精选推荐。
📋 工具概览

memgraph-ingester 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。

GitHub Stars
⭐ 6
开发语言
Java
支持平台
Windows / macOS / Linux / Android
维护状态
轻量级项目,按需更新
开源协议
MIT
AI 综合评分
7.1 分
工具类型
Agent工作流
Forks
📖 中文文档
以下内容由 AI Skill Hub 根据项目信息自动整理,如需查看完整原始文档请访问底部「原始来源」。

memgraph-ingester 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。

📌 核心特色
  • 可视化 Agent 工作流编排,无需编写复杂代码
  • 支持多步骤自动化任务链,实现全流程无人值守
  • 与外部 API、数据库和第三方服务无缝集成
  • 内置错误处理与自动重试机制,保障稳定运行
  • 提供可复用的自动化模板,快速在同类场景部署
🎯 主要使用场景
  • 自动化日常重复性工作,将精力集中于创造性任务
  • 构建数据采集 → 处理 → 输出的完整自动化管线
  • 实现跨平台、跨系统的数据流转和业务协同
以下安装命令基于项目开发语言和类型自动生成,实际以官方 README 为准。
安装命令
# 克隆仓库
git clone https://github.com/ousatov-ua/memgraph-ingester
cd memgraph-ingester

# 查看安装说明
cat README.md

# 按 README 完成环境依赖安装后即可使用
📋 安装步骤说明
  1. 访问 GitHub 仓库获取工作流文件
  2. 在对应平台(Dify / Flowise / Make 等)中找到「导入工作流」功能
  3. 上传工作流文件
  4. 按照提示配置必要的环境变量和 API Key
  5. 运行测试确认流程正常后投入使用
以下用法示例由 AI Skill Hub 整理,涵盖最常见的使用场景。
常用命令 / 代码示例
# 查看帮助
memgraph-ingester --help

# 基本运行
memgraph-ingester [options] <input>

# 详细使用说明请查阅文档
# https://github.com/ousatov-ua/memgraph-ingester
以下配置示例基于典型使用场景生成,具体参数请参照官方文档调整。
配置示例
# memgraph-ingester 配置说明
# 查看配置选项
memgraph-ingester --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 MEMGRAPH_INGESTER_CONFIG="/path/to/config.yml"
📑 README 深度解析 真实文档 完整度 50/100 查看 GitHub 原文 →
以下内容由系统直接从 GitHub README 解析整理,保留代码块、表格与列表结构。

Memgraph Ingester

Build Maven Central License: MIT Visitors GitHub commits GitHub last commit

memgraph-ingester-readme-banner-640x320.svg

Memgraph Ingester indexes a Java or JavaScript/TypeScript codebase into Memgraph so AI agents can query a real code graph instead of repeatedly scanning raw files.

Use it when you want your agent to quickly answer questions such as:

  • What classes, methods, files, and packages exist?
  • What extends or implements this type?
  • Who calls this method?
  • What durable project rules, decisions, findings, risks, and tasks should the agent remember?

The normal path is simple:

  1. Run Memgraph.
  2. Download one ingester executable.
  3. Run one command for Java or one command for JS/TS.
  4. Optionally connect your AI agent through MCP or mgconsole.

No source code is uploaded by the ingester. It reads local files, writes graph nodes to your Memgraph instance over Bolt, and exits.

Requirements

For normal use:

  • Memgraph, either local, Docker, or an existing Bolt endpoint.
  • One ingester download from the release page.

Runtime requirements by artifact:

ArtifactJava required?Node.js required for JS/TS?
Native executableNoNo, managed mode handles it
Shaded JARJava 25 JRENo, managed mode handles it

Optional tools:

  • mgconsole, if you want to query Memgraph directly without MCP (produces much fewer tokens)
  • Maven, only if you want a richer Java classpath or you want to build from source.
  • Java 25 SDK, only if you build from source.

Use as a Maven dependency

<dependency>
  <groupId>io.github.ousatov-ua</groupId>
  <artifactId>memgraph-ingester</artifactId>
  <version>9.0.3</version>
</dependency>

Agent Setup

The graph is useful directly, but it becomes much more powerful when your agent is told to use it. The executable can write project-scoped Memgraph instructions to the agent's local instruction file. It replaces its own previously managed block when one already exists, so rerunning the command keeps AGENTS.md, CLAUDE.md, or another instruction file tidy.

Run the command from the repo you just ingested, using the same --project value.

Code graph guidance is installed by default:

memgraph-ingester --init-instructions -P my-project
memgraph-ingester -P my-project --instructions-agent codex

Add optional Memory workflow instructions when you want agents to create and maintain durable Memgraph Memories:

memgraph-ingester --init-instructions -P my-project --with-memories
memgraph-ingester -P my-project --instructions-agent codex --with-memories

Agent presets choose the default target file:

memgraph-ingester --init-instructions -P my-project --instructions-agent codex
memgraph-ingester --init-instructions -P my-project --instructions-agent claude
memgraph-ingester --init-instructions -P my-project --instructions-agent gemini
memgraph-ingester --init-instructions -P my-project --instructions-agent github

When --instructions-agent is present, --init-instructions is optional.

Use --instructions-file to target a specific file:

memgraph-ingester -P my-project --instructions-file .github/copilot-instructions.md

The legacy helper scripts now delegate to the executable and accept the same optional flags:

MEMGRAPH_INGESTER_BIN=./memgraph-ingester-macos-arm64 script/init-memgraph-codex.sh my-project --with-memories

Commit the updated instruction file so future agent sessions get the same graph guidance.

For Enthusiasts: Build It Yourself

You do not need this section to use the tool. Use the release downloads unless you want to hack on the project.

Build the shaded JAR

Requirements:

  • Java 25 SDK.
  • Maven 3.9+.
git clone https://github.com/ousatov-ua/memgraph-ingester.git
cd memgraph-ingester
mvn clean package -Pshade -DskipTests

Output:

target/memgraph-ingester.jar

Run it:

java -jar target/memgraph-ingester.jar --help

Build a native executable

Native builds use GraalVM Native Image and build for the OS where Maven runs.

mvn clean package -Pnative-macos -DskipTests
mvn clean package -Pnative-linux -DskipTests
mvn clean package -Pnative-windows -DskipTests

Use the profile that matches your operating system.

Quick Start

Java Guide

The Java adapter reads .java files using JavaParser with Java 25 syntax support. It should handle most earlier Java versions as well.

Captured Java structure:

  • Packages and files.
  • Classes, interfaces, annotations, enums, records, and nested classes.
  • Methods, constructors, fields, visibility, return types, static flags, line ranges, and synthetic flags.
  • EXTENDS, IMPLEMENTS, DECLARES, DEFINES, CONTAINS, ANNOTATED_WITH, and best-effort CALLS.

Use --classpath whenever you can. Without dependency JARs, the ingester still works, but external types may fall back to simple names and some call edges may be missing.

For Maven projects:

CP=$(mvn -q dependency:build-classpath -DincludeScope=test -Dmdep.outputFile=/dev/stdout 2>/dev/null)

Use -DincludeScope=test when you ingest tests or test fixtures. It includes JUnit, Testcontainers, mocking libraries, and other test-scoped dependencies.

Generated code is indexed only if you ingest it:

<ingester> \
  --source target/generated-sources/annotations \
  --bolt bolt://localhost:7687 \
  --project my-java-project

Do not pass --wipe-project-code on the generated-source pass unless you want to replace the previous graph with generated sources only.

JavaScript and TypeScript Guide

Use --language js, --language javascript, --language ts, or --language typescript.

Accepted source extensions:

  • .js
  • .jsx
  • .ts
  • .tsx
  • .mts
  • .cts
  • .d.ts
  • .d.mts
  • .d.cts
  • .mjs
  • .cjs

Only source files under --source are considered. Use the repository root as --source when you want root JavaScript config files or support scripts indexed alongside application source. tsconfig.json and configs from its extends chain are read for TypeScript path aliases when present, but they are not indexed as code nodes.

Skipped paths:

  • Anything under node_modules.

Captured JS/TS structure:

- Files and synthetic module owners. - Classes and class expressions assigned to variables. - Interfaces and type aliases as graph interfaces. - Class/interface EXTENDS and class IMPLEMENTS relationships, including relative imports and tsconfig.json path aliases, including those inherited from extended configs, that resolve under --source. - Interface and object-literal type members as :Field/:Method declarations. - TypeScript enums as graph classes with isEnum = true and kind = "enum", with enum members as :Field declarations using kind = "enum-member". - Top-level functions and variables under the module owner. - Exported callable aliases and class re-export aliases as graph-visible declarations for their public export names. - Methods, constructors, function-valued class fields, fields, abstract class metadata, bodyless abstract/optional method signatures, static flags, line ranges, and kinds. - Decorators as annotations, preserving namespace-qualified decorator FQNs when possible. - Angular decorators with framework metadata when detected. - Syntax-based best-effort call edges, including top-level IIFEs/callbacks, local function constructors, and deferred resolution for resolvable relative imports. - Relative import and tsconfig.json path-alias resolution, including extended configs, prefers TypeScript source files over emitted JavaScript when both exist for the same local module path.

Runtime modes:

ModeUse whenNetwork
managedYou want the ingester to own the parser runtime. This is the default.Downloads once if cache is missing.
systemYou want to use node from PATH.No Node download. TypeScript may still be managed unless already cached.
offlineYou want no downloads and the cache is already warm.No downloads. Fails if cache is missing.

Custom cache:

<ingester> \
  --source . \
  --bolt bolt://localhost:7687 \
  --project my-js-project \
  --language js \
  --js-runtime-cache /path/to/cache \
  --wipe-project-code

Custom pinned versions:

<ingester> \
  --source . \
  --bolt bolt://localhost:7687 \
  --project my-js-project \
  --language js \
  --js-node-version 22.11.0 \
  --js-typescript-version 5.6.3 \
  --wipe-project-code

JS/TS caveat: JavaScript is dynamic. Dynamic dispatch, dependency injection, monkey-patching, framework templates, and generated code can produce missing call edges. A missing JS/TS CALLS edge does not prove a call never happens.

CLI Reference

Exit codes:

CodeMeaning
0Success
1Invalid arguments or runtime setup failure
2One or more files failed to parse or ingest

Options:

OptionShortRequiredDefaultDescription
--source-syesRoot directory to scan.
--bolt-byesMemgraph Bolt URL, for example bolt://localhost:7687.
--project-PyesLogical project name. Namespaces all graph nodes.
--user-unoemptyMemgraph username.
--pass-pnoemptyMemgraph password.
--threads-tno1Parser threads. Each thread gets its own Bolt session.
--wipe-project-codenofalseDelete this project's code graph before ingesting.
--wipe-project-memoriesnofalseDelete this project's memory graph before ingesting.
--apply-schemanofalseApply Memgraph constraints and indexes before ingesting.
--wipe-allnofalseDelete all data from Memgraph.
--incrementalnofalseSkip files whose last-modified timestamp matches the graph.
--watch-wnofalseWatch the source directory and re-ingest changes.
--classpathnoemptyPlatform-separated JAR paths for Java symbol resolution.
--languagenojavajava, js, javascript, ts, or typescript.
--js-runtime-modenomanagedmanaged, system, or offline.
--js-runtime-cacheno~/.cache/memgraph-ingesterCache directory for managed Node.js and TypeScript downloads.
--js-node-versionno22.11.0Pinned Node.js version for managed JS/TS parsing.
--js-typescript-versionno5.6.3Pinned TypeScript compiler package version. A leading v is accepted.
--check-js-runtimenofalseRun a local JS runtime smoke check without connecting to Memgraph.
--init-instructionsnofalseWrite or replace managed agent instructions and exit. Includes Code guidance by default.
--instructions-agentnocodexAgent preset: codex, claude, gemini, github, or copilot. Implies --init-instructions when explicitly provided.
--instructions-filenopreset fileInstruction file to update. Overrides --instructions-agent and implies --init-instructions.
--with-memoriesnofalseInclude optional Memory workflow instructions when initializing agents.
--helpnoPrint CLI help.
--versionnoPrint CLI version.

Parallel ingestion:

ThreadsTypical speedupBottleneck
11xSequential parse and write
4about 2.5x to 3xWrite serialization starts
8about 3x to 4xDiminishing returns
16+about 3x to 4xWrites saturated

Use 4 to 8 threads for large projects on most machines. Values higher than your CPU core count rarely help.

📚 实用指南(长尾问题)
适合谁
  • 需要让 Claude / Cursor 操作本地工具的 AI 工程师
  • 构建多智能体协作系统的 Agent 开发者
最佳实践
  • 配置 MCP 服务器时建议使用 stdio 传输 + JSON-RPC,避免暴露公网
  • 生产部署优先使用 Docker Compose 隔离依赖,并挂载 volume 持久化数据
  • Agent 任务先做 dry-run 验证工具调用链,再开启自主执行
常见错误
  • API key 直接提交到 git 仓库(请用 .env 并加入 .gitignore)
  • MCP 配置路径拼错或权限不足,重启 Claude Desktop 才生效
  • 容器内无法访问宿主机 localhost — 使用 host.docker.internal
部署方案
  • Docker:memgraph-ingester 提供官方镜像,docker compose up 一键启动
  • CLI:直接 npm install -g / pip install,命令行调用
  • 云端托管:可放在 Vercel / Railway / Fly.io 等 PaaS 平台
相关搜索
memgraph-ingester 中文教程memgraph-ingester 安装报错怎么办memgraph-ingester MCP 配置memgraph-ingester Docker 部署memgraph-ingester Agent 工作流memgraph-ingester 与同类工具对比memgraph-ingester 最佳实践memgraph-ingester 适合谁用
⚡ 核心功能
👥 适合人群
自动化工程师和运维人员项目经理和业务分析师希望减少重复性工作的专业人士数字化转型团队
🎯 使用场景
  • 自动化日常重复性工作,将精力集中于创造性任务
  • 构建数据采集 → 处理 → 输出的完整自动化管线
  • 实现跨平台、跨系统的数据流转和业务协同
⚖️ 优点与不足
✅ 优点
  • +MIT 协议,可免费商用
  • +大幅减少重复性人工操作
  • +可视化流程,清晰直观
  • +可扩展性强,支持复杂场景
⚠️ 不足
  • 初始配置和调试需投入一定时间
  • 强依赖外部服务的稳定性
  • 复杂场景需具备一定技术基础
⚠️ 使用须知

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。

📄 License 说明

✅ 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。

🔗 相关工具推荐
❓ 常见问题 FAQ
memgraph-ingester 是一款Java开发的AI辅助工具。开源AI工作流:Ingester of Java structure in Memgraph. Speed up your AI agent!。⭐6 · Java 主要应用场景包括:编程开发。
💡 AI Skill Hub 点评

AI Skill Hub 点评:memgraph-ingester 的核心功能完整,质量良好。对于自动化工程师和运维人员来说,这是一个值得纳入个人工具库的选择。建议先在非生产环境试用,再逐步推广。

⬇️ 获取与下载
⬇ 下载源码 ZIP

✅ MIT 协议 · 可免费商用 · 直接从 aiskill88 服务器下载,无需跳转 GitHub

📚 深入学习 memgraph-ingester
查看分步骤安装教程和完整使用指南,快速上手这款工具
🌐 原始信息
原始名称 memgraph-ingester
原始描述 开源AI工作流:Ingester of Java structure in Memgraph. Speed up your AI agent!。⭐6 · Java
Topics workflowaiai-agentai-agentsclaude-codecodexjava
GitHub https://github.com/ousatov-ua/memgraph-ingester
License MIT
语言 Java
🔗 原始来源
🐙 GitHub 仓库  https://github.com/ousatov-ua/memgraph-ingester

收录时间:2026-05-22 · 更新时间:2026-05-22 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。