MinerU
Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
Read authorized applications and task status from an AI Native System deployment.
Skills that teach the AI this kind of work; they work alongside this one
Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
A privacy-first app that strips AI watermarks from content you own.
Agent skills for Obsidian. Teach your agent to use Obsidian CLI and open formats including Markdown, Bases, JSON Canvas.
Lists are ordered by a usage score (out of 100) built from these parts
| Popularity | 4/80 | GitHub stars or weekly downloads: the larger counts 60, the smaller 20 |
|---|---|---|
| Freshness | 12/12 | Full marks for a release or code change within 30 days |
| Still answering | 0/8 | The cloud endpoint answered at the last check |
| Usage score | 16/100 |
Quality: Passed 2/3 basic checks (shown, not used for ordering) · Momentum: No data yet (Used only on the Trending tab)
Pick the AI app you use, then copy the settings.
Passed 2/3 checks · Runs on your computer
First: which AI do you use?
Pick your app and the steps appear below.
Works with Claude, with some setup first
No key, nothing to install · Only the Claude desktop app (Windows or Mac), not the web or phone
Download the maker’s .mcpb file
The .mcpb file comes straight from the maker; we do not change it
Use AI Native System and tell me what you can do for me, then show me one thing✓ If Claude says it is using AI Native System or shows a AI Native System tool in the answer, you are connected
Menu names checked against Anthropic’s help page on Nov 11, 2026 · your screen may differ slightly by app version
ChatGPT cannot connect AI Native System yet
AI Native System runs on your computer, but ChatGPT can only connect servers that run on the web
A · Use an app that can connect AI Native System
Claude Code cannot connect AI Native System yet
The maker only provides an installer for the Claude desktop app
A · Use an app that can connect AI Native System
Cursor cannot connect AI Native System yet
The maker only provides an installer for the Claude desktop app
A · Use an app that can connect AI Native System
VS Code cannot connect AI Native System yet
The maker only provides an installer for the Claude desktop app
A · Use an app that can connect AI Native System
Other apps that use MCP, such as Windsurf, Zed or Gemini CLI, each add servers differently. See how in the app list.
JSON settings are under “For developers” below
Based on how this server connects and on each app’s own docs, checked Nov 10, 2026.
Run on your computer in
ClaudeGooseLM StudioMsty StudioCherry StudioVisual StudioWindsurf (Devin Desktop)ZedKiroClineContinueJetBrains AI AssistantCodex CLIGemini CLIOpenCodenodecodeWarpLibreChatRaycast
Runs on your computer; the tool list appears in your AI app after you install it.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 20, 2026
The registry name was verified with the GitHub account dengyier, which owns the repo
Code last changed 3 days ago
GitHub could not identify the licence type; open the file to check
This one runs on your computer; there is no internet endpoint
Runs on your computer and uses local settings instead
Runs on your computer; we can only read tool lists from internet endpoints
Not checked: only npm and PyPI packages with a version can be checked
If you made this, paste this code into your README. The badge links back to this page.
[](https://mcp.popnx.com/en/servers/dengyier-ai-native-system)Ask, share a tip or report that it does not work. Sorted by what people marked helpful.
No posts yet. Be the first to ask or share.
Posts here are written by users, not the PopNx team.
No reviews yet
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 2 of 3 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).
The GitHub repo last changed 3 days ago, and the latest version (0.1.1) reached the registry 3 days ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need MCPB on your computer.
In the same category: Puter MCP, XcodeBuildMCP, Vexa, Atlassian Rovo MCP Server, Anki MCP Server
Same category, other makers, commands already checked. Official ones first, then by stars.
Reading a card:↓ Downloads per weeki★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
Puter MCP enables AI tools to interact with Puter: manage files, websites, workers, and more
XcodeBuildMCP provides tools for Xcode project management, simulator management, and app utilities.
Meeting bot and transcripts for Google Meet, Teams and Zoom. Live or after, speakers labelled.
Atlassian’s official server lets AI search and manage Jira work and Confluence pages. Connects to Atlassian’s cloud with OAuth sign-in.
MCP server for Anki flashcards: adaptive review, notes, media, and deck management via AnkiConnect.
| This one AI Native System | Puter MCP | XcodeBuildMCP | Vexa | Atlassian Rovo MCP Server | Anki MCP Server | |
|---|---|---|---|---|---|---|
| Runs where | Installs on your computer | Runs in the cloud, nothing to install | Installs on your computer | Runs in the cloud, nothing to install | Runs in the cloud, nothing to install | Installs on your computer |
| Needs a key | No | No | No | Yes | No | No |
| Add with + in Claude | No | No | No | No | Yes | No |
| Basic checks | 2/3 | 4/4 | 3/3 | 2/4 | 4/4 | 3/3 |
| License | NOASSERTION | AGPL-3.0 | MIT | Apache-2.0 | Apache-2.0 | MIT |
| Price | No cost listed in the registry | No cost listed in the registry | No cost listed in the registry | No cost listed in the registry | No cost listed in the registry | No cost listed in the registry |
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
面向企业业务的 AI 原生框架,用于构建可审批、可验证、可恢复的 Agent 工作流。应用声明业务对象、关系、规则和动作,Agent 根据获准的事实和知识提出建议。涉及业务写入时,可信执行服务核对审批、OpenWorkProof 授权和源系统版本,再执行条件写入并核验签名回执。
六层架构 · 快速开始 · 接入应用 · 已实现与待完成清单
| 层 | 负责什么 |
|---|---|
| 1. 基础设施 | PostgreSQL 持久化业务状态、发布、审计与检查点。采用行级租户隔离,并提供部署、观测和恢复演练入口 |
| 2. 数据 | Ontology 声明对象和关系。事实同步记录来源、版本、冲突与历史。文档经 LlamaIndex 分段,在 PostgreSQL 和 pgvector 中进行权限内的中文混合检索 |
| 3. 模型 | LiteLLM 路由生成和 embedding,记录调用用量与成本,并通过 Langfuse 上报脱敏观测数据 |
| 4. AI 中台核心 | Agent Runtime 和 LangGraph 编排阶段、审批与恢复。OpenWorkProof 绑定动作授权,可信执行服务通过 MCP 或 HTTP 执行并对账 |
| 5. AI 应用 | Python SDK、应用 API 和浏览器运行台承载问答、事实核查、审批、任务和证据 |
| 6. 企业业务应用 | 通过连接器读取业务系统的权威对象、关系和版本。获准的业务动作以条件写入、源端签名回执和可核验终态闭环 |
Agent 进程只持模型侧凭据,从可信服务收到按权限裁剪的上下文。业务凭据由执行服务持有。已批准动作在派发前复核当前事实、资源权限与签名,源端结果不确定时按原操作 ID 查询回执。 可编辑架构图 · Agent 隔离及恢复演练
python -m pip install ainsys
ains --help
AINS 短包 ainsys 安装同版本的 ai-native-system 主框架,并提供 ains 和 ains-mcp 命令。 AINS 短包 · 主框架安装包 · 短包源码
在仓库根目录使用 Python 3.11–3.15。命令使用内置模型和合成配置,在临时 SQLite 数据库中运行,无需下载模型或提供密钥。
ai_native_tmp=$(mktemp -d)
python3 -B -m ai_native_system --db "$ai_native_tmp/framework.sqlite" demo
| 演示 | 可观察结果 |
|---|---|
single | 单 Agent 读取对象、引用规则提出建议。模拟审批后,对象 asset-1 由 pending 变为 verified |
collaboration | 两个 Agent 阶段通过结构化记录交接。模拟审批后,artifact-1 按关联对象状态由 draft 变为 published |
输出包含阶段状态、审批、事件和动作回执。日常 CLI 帮助可用 python3 -m ai_native_system --help 查看。
准备 uv、可用的 Docker Engine 和 Docker Compose。以下脚本创建独立临时测试库。实测环境使用 Python 3.12、PostgreSQL 18.7 和 pgvector 0.8.7。
uv sync --locked --all-extras --python 3.12 --no-build
uv run --locked --all-extras --python 3.12 python -B deployment/local-postgres/verify.py uv run --locked --all-extras --python 3.12 python -B deployment/local-postgres/verify_p7.py uv run --locked --all-extras --python 3.12 python -B deployment/local-postgres/verify_p8.py --only applications
第一个验证脚本检查连接池、事务、角色和向量距离。P7 覆盖文档入库、索引与文档权限、中文混合检索、引用和记忆。P8 验证参考应用沿用权限、检索与可信执行。 环境与锁定依赖 · P7 验收 · P8 验收
--registry 加载。主框架的 wheel 可从 GitHub 发布页 下载。安装后用 ai-native-system --db demo.sqlite demo 运行本地演示。接入 PostgreSQL 的环境可安装 wheel 的 postgres 和 integrations 可选依赖。
python -m pip install "ainsys[mcp]"
用 MCP 客户端以 stdio 启动 ains-mcp,并提供部署端 API 地址 AI_NATIVE_API_URL 与访问令牌 AI_NATIVE_API_TOKEN。服务端提供 list_applications 和 get_task 两种只读工具,使用当前 API 身份的授权。远端 API 使用 HTTPS,本机回环允许 HTTP。
支持 MCPB 0.4 与 UV 的客户端可安装 AI Native System MCPB,并按提示填入 API 地址和令牌。 官方 MCP Registry 条目
主框架及 MCP 服务端适用根目录的非商业使用许可。独立 Python SDK 和企业动作接口定义采用 MIT。 主框架与 MCP 许可 · MIT SDK
数据库管理员先在独立测试库启用 public schema 下的 pgvector 扩展,并将 ai_native schema 的所有权交给迁移角色。通过环境或密钥服务提供 AI_NATIVE_POSTGRES_MIGRATION_DSN 和最小权限的 AI_NATIVE_POSTGRES_DSN。下面命令使用已安装集成依赖的 Python 环境,模型配置通过 model_config 指向本地文件。
python -m ai_native_system --postgres migrate
python -m ai_native_system --postgres --registry fixtures/p8-registry.json \ --models "$model_config" serve --application-catalog fixtures/p8-applications.json
完成企业身份、权限和发布配置后,在本机访问 http://127.0.0.1:8789/console/。实际连接信息和密钥通过部署环境提供。 主库与迁移 · 身份、委派和发布 · 模型配置样例
为请求主体授予实际知识索引的 read / knowledge-index / 索引ID 权限及当前文档的读取权限。执行迁移后,拥有词法索引读取权限的服务可用 --method hybrid 查询。问答及摘编应用可在目录条目加入 "retrieval_method": "hybrid"。默认配置使用精确向量检索。
python -m ai_native_system --postgres --tenant "$tenant" --models "$model_config" \
p7 search '“保留审批记录”' --method hybrid --actor "$actor"
中文词组与向量候选在同一授权范围内融合,完整条款号和带引号的连续短语支持精确匹配,返回正文前复核权限与原文引用。 索引授权与启用步骤
需要稳定复跑时,在应用发布草稿中显式固定知识索引和检索方法。用 release select 切换发布或回到仍有效的旧发布,系统在移动指针前校验目标索引的文档覆盖和模型配置。 固定发布与回滚
| 模块 | 作用 |
|---|---|
| Ontology · Registry | 声明领域类型、关系、规则、应用与动作 |
| Runtime · 图编排 | 管理 Agent 阶段、结构化交接、人工审批和持久恢复 |
| 知识检索 · 中文词法查询 | 在同一索引与权限范围内召回并复验原文 |
| OpenWorkProof 执行校验 · 动作与回执契约 | 绑定审批、事实、权限和源端证据 |
| 应用 API · Python SDK · 运行台 | 访问获准的应用、任务、事实、审批与证据 |
Python 客户端可独立安装,通过 from ai_native_sdk import Client 连接已部署的应用接口。下载包、用法和只读示例见 SDK 文档。
在本地手动推进一个任务
ai_native_tmp=$(mktemp -d)
ai_native_db="$ai_native_tmp/manual.sqlite" python3 -m ai_native_system --db "$ai_native_db" init
task_id=$(python3 -m ai_native_system --db "$ai_native_db" \ submit single asset-1 --actor requester --request-key manual-1 | python3 -c 'import json, sys; print(json.load(sys.stdin)["id"])')
python3 -m ai_native_system --db "$ai_native_db" advance "$task_id" python3 -m ai_native_system --db "$ai_native_db" decide "$task_id" approve --actor reviewer python3 -m ai_native_system --db "$ai_native_db" advance "$task_id" python3 -m ai_native_system --db "$ai_native_db" inspect "$task_id"
在独立的演示任务上,执行阶段可使用 advance … --simulate-timeout before-commit 或 after-commit 分别模拟提交前超时与提交后确认丢失。后续使用同一任务 ID 推进或查看,核对对象版本与回执。
可选的本机模型接入
设置 AI_NATIVE_MODEL_BASE_URL、AI_NATIVE_MODEL_NAME 和服务需要的 AI_NATIVE_MODEL_API_KEY,为本地演示接入支持 Chat Completions JSON 输出的模型。若已有 GGUF 模型与 node-llama-cpp,可先启动本机回环桥接。
AI_NATIVE_LLAMA_RUNTIME_DIR="/path/to/runtime" \
AI_NATIVE_GGUF_MODEL_PATH="/path/to/model.gguf" \ node local_model.mjs
将模型 URL 设为 http://127.0.0.1:8798/v1,模型名使用 GGUF 文件去掉扩展名的名称。企业模型路由、费用预算、观测和回退配置见 P5 模型接入验收。
Copyright © 2026 dengyier. 原创代码、文档及架构图的知识产权由 dengyier 持有。主框架采用 AI Native System 非商业使用许可 v1.3。
主框架的个人学习和非商业使用免费。将主框架用于企业生产、客户交付、收费托管等商业用途,须事先获得单独书面授权。GitHub 平台内的查看与 Fork 适用平台服务条款,对外分发和再许可依项目许可执行。 商业授权联系 dengyier
由 dengyier 上传到 PyPI 的安装包可依许可第 4 条原样免费分发和镜像。
独立 Python SDK 和接入示例,以及企业动作接口定义及配套说明,可按 MIT 许可使用。 SDK MIT 许可 · 接口定义 MIT 许可
第三方依赖适用各自的上游许可证。 依赖与许可证