登录

openagent

Haohao-endMIT活跃维护

把 OpenAI Deep Research 的深度推理和 Dify 的可视化工作流合到一起的垂直 AI Agent 构建平台，支持 RAG 与 A2A 委派。

What if OpenAI Deep Research and Dify were one platform? OpenAgent — harness architecture for rapidly building vertical AI agents, with deep reasoning loops, visual workflows, RAG, and A2A delegation.

v1.1.4发布于 2026-06-15 08:54 UTC753star +5 / 7 天78fork同步于 2026-11-10 17:56 UTC
Agent 框架深度推理可视化工作流RAGA2A 委派
对比 GitHub 官网agentaicelerydeepagentsdeepresearch

怎么上手

由 README 提炼，执行前请确认
原文 
git clone https://github.com/Haohao-end/openagent.git && cd openagent
cp api/.env.example api/.env
cd docker && docker compose up -d --build
硬件要求按 README 说明

README 未说明 GPU 或显存要求；全栈推荐 8 GB 以上内存，需 Docker 20.10+ 与 Docker Compose 2.x，并至少准备一个模型厂商的 API Key。

有 Docker 镜像
下载：源码 zip·GitHub Releases（安装包）本站不镜像安装包

项目说明

由 README 自动翻译提炼
没人工点评？写下你的经验

开源全栈 AI Agent 平台，把 Deep Research 式深度推理与 Dify 式可视化编排合为一体

把 OpenAI Deep Research 的深度推理和 Dify 的可视化工作流合到一起的垂直 AI Agent 构建平台，支持 RAG 与 A2A 委派。

核心能力

  • 用可视化工作流编排 LLM、工具、检索、代码与分支节点
  • 开启 Deep Research 深度思考模式拆解复杂任务并串联能力
  • 通过 A2A 把用户请求路由到已发布的公共 Agent
  • 管理数据集与文档分段，接入 RAG 混合检索
  • 发布应用后经 OpenAPI 以 REST/SSE 方式调用

容易踩的坑

启动前必须在 api/.env 中填好 JWT_SECRET_KEY、Postgres/Redis/Weaviate 密码及至少一个模型厂商 Key，否则服务起不来；默认端口为前端 3000、API 5001、Nginx 80，注意端口占用；本地开发可只起 db/redis/weaviate 三个容器，API 与 UI 需自行在本机运行。

内容由模型根据 README 生成，可能不完整或不准确，以官方文档为准。看原文 README

最近发版

5 个版本v1.1.4
全部 release
  1. v1.1.4最新正式版2026-06-15 · 2026-06-15 08:54 UTC

    This release primarily polishes the README and project presentation, including sponsor display updates (now featuring both Atlas Cloud and Bloome), acknowledgment for Ryan B.'s security contribution, multiple UI fixes, and improvements to the development workflow. It also fixes several issues related to the Sidebar, Store, Agent binding, and dynamic loading to further enhance workspace consistency and reliability.

    It is recommended for all users running the v1.1.3 stack in local development, Docker Compose, or self-hosted environments.

    Highlights

    README & Project Presentation Improvements

    • Updated sponsor section to prominently feature Atlas Cloud and Bloome.
    • Refactored provider ecosystem titles and improved sponsor block structure.
    • Added acknowledgment to Ryan B. for SSRF vulnerability reporting.
    • Optimized ER diagram layout for the upload_file and process_rule tables.
    • Added *.Zone.Identifier to .gitignore (better Windows user experience).

    UI & Functional Fixes

    • Fixed duplicate entries in Sidebar Store and improved icon differentiation for MCP/Workflow/Skills.
    • Optimized Agent binding logic to runtime tools and restored proper App layout behavior.
    • Added support for dynamic loading of Recent Conversations.
    • Removed duplicate MCP entries and hardened release tags.
    • Added error handling for the A2A conversation messages endpoint.
    • Force-tracked safe_http_client.py and cleaned up related ignores.
    • Removed outdated Atlas Cloud gpt-5.2 entry while maintaining compatibility.

    Deployment & Developer Experience

    • Improved Skills catalog and Store UI synchronization.
    • Enhanced overall code and documentation consistency.

    Developer notes

    • When modifying the README, keep sponsor information synchronized between English and Chinese versions.
    • New UI strings should continue to be managed through i18n keys (following v1.1.3 conventions).
    • Before publishing or deploying, run the frontend validation flow: 
      cd ui
npm run type-check
npm run lint
npm run build
  2. v1.1.32026-06-07 · 2026-06-07 14:28 UTC
    查看更新说明

    Overview

    OpenAgent v1.1.3 is a focused patch release after v1.1.2.

    This release introduces the first structured internationalization (i18n) layer for the OpenAgent frontend. It prepares the product workspace for a better English and Chinese experience, centralizes user-facing copy into locale resources, reduces hard-coded UI text, and improves consistency across common product flows.

    It also includes release-cycle stability work around frontend build readiness, Skills catalog packaging, API startup catalog synchronization, and Alembic migration handling for deployment environments.

    It is recommended for users running the v1.1.2 stack in local development, Docker Compose, or self-hosted environments.

    Highlights

    Internationalization foundation

    • Added frontend i18n support for the OpenAgent Vue workspace.
    • Introduced locale resource organization for user-facing product copy.
    • Added bilingual localization groundwork for English and Chinese UI experiences.
    • Centralized common labels, actions, placeholders, navigation text, messages, and page copy.
    • Prepared frontend pages and components to resolve display text through i18n keys instead of hard-coded strings.
    • Added safer fallback behavior so untranslated or missing keys can degrade more gracefully.

    Bilingual workspace experience

    • Improved localized copy coverage across core product areas:

      • Home assistant
      • App workspace
      • Workflow builder
      • Dataset and document management
      • Public apps, tools, and workflow browsing
      • Authentication, settings, dialogs, and common UI states

    • Reduced mixed-language UI text in the frontend workspace.

    • Improved consistency for page titles, menu items, buttons, empty states, validation messages, and toast notifications.

    • Prepared the UI for future language expansion beyond the initial English and Chinese coverage.

    Language runtime behavior

    • Added runtime language resolution support for the frontend application.
    • Prepared the app bootstrap flow so localized pages can render through the i18n layer.
    • Added support for preserving or reusing the active language during the frontend session.
    • Kept the default English experience available while enabling Chinese UI usage.

    Release and deployment polish

    • Ensured localization resources are included in the frontend build path.
    • Kept Vite and Docker build behavior compatible with the new i18n structure.
    • Continued release stability cleanup around built-in Skills catalog packaging.
    • Improved API startup behavior by syncing the Skills catalog during service initialization.
    • Fixed catalog markdown tracking so packaged Skills documentation is included reliably.
    • Resolved Alembic migration head split issues that could affect deployment upgrades.

    Developer notes

    • New user-facing frontend strings should be added through locale keys instead of being hard-coded in Vue components.

    • When adding or changing UI copy, update both English and Chinese locale resources in the same change wh

  3. v1.1.22026-05-26 · 2026-05-26 15:30 UTC
    查看更新说明

    Overview

    OpenAgent v1.1.2 is a focused patch release after v1.1.1.

    This release expands the provider ecosystem with Atlas Cloud, improves the App Workspace and Deep Research experience, strengthens Assistant MCP runtime behavior, packages the built-in Skills catalog more reliably, and fixes deployment issues around Alembic migration heads and Docker startup catalog synchronization.

    It is recommended for users running the v1.1.1 stack in local development, Docker Compose, or self-hosted environments.

    Highlights

    Atlas Cloud provider integration

    • Added Atlas Cloud as an OpenAI-compatible model provider.

    • Added a backend atlascloud provider to the language model registry.

    • Implemented an OpenAI-compatible Atlas Cloud chat adapter using the existing ChatOpenAI integration path.

    • Added environment-based Atlas Cloud credential and endpoint resolution:

      • ATLASCLOUD_API_KEY
      • ATLASCLOUD_API_BASE
      • fallback support for ATLAS_CLOUD_API_KEY
      • fallback support for ATLAS_CLOUD_API_BASE

    • Added the default Atlas Cloud API base URL:

      https://api.atlascloud.ai/v1
  4. v1.1.12026-05-05 · 2026-05-05 16:40 UTC
    查看更新说明

    Overview

    OpenAgent v1.1.1 is a focused patch release after v1.1.0.

    This release improves the reliability of the home assistant experience, public app and public workflow behavior, provider/model metadata, runtime timeout defaults, health checks, Docker deployment defaults, and README/brand presentation.

    It is recommended for users running the v1.1.0 stack in local development, Docker Compose, or self-hosted environments.

    Highlights

    Home assistant and new conversation flow

    • Added a more reliable new conversation flow for the home assistant.
    • The sidebar Home entry now starts a fresh conversation flow instead of only performing a plain route jump.
    • Added support for opening a fresh home conversation through the new_conversation query parameter.
    • Improved frontend conversation reset behavior to avoid stale state when starting a new assistant session.

    Public app and workflow reliability

    • Improved public app A2A message handling for more stable long-running responses.
    • Improved public workflow listing and tag filtering behavior.
    • Moved more filtering work closer to the database query layer to reduce unstable post-filtering paths.
    • Added safer fallback behavior when workflow tags are missing or incomplete.

    Provider runtime and model metadata

    • Normalized model/provider pricing display to RMB.
    • Added and refreshed provider metadata for supported model integrations.
    • Added DeepSeek v4 Flash and DeepSeek v4 Pro model definitions.
    • Added safer default request timeout behavior across compatible model clients.
    • Improved consistency across OpenAI-compatible provider configurations.

    UI and message rendering

    • Improved home page message rendering for long responses.
    • Improved Markdown layout behavior in assistant and user message bubbles.
    • Reduced overflow issues for long text, code blocks, tables, and structured content.
    • Improved frontend navigation behavior around the home workspace.

    Deployment and runtime configuration

    • Added a lightweight /healthz endpoint for container and reverse proxy health checks.
    • Updated Docker Compose defaults for safer local/self-hosted deployment.
    • Added API healthcheck support.
    • Improved default environment configuration for Flask, Gunicorn, SQLAlchemy, Redis, Weaviate, and LLM request timeout settings.
    • Updated api/.env.example with more complete runtime configuration references.

    Documentation and brand updates

    • Refreshed README and README_ZH content.
    • Added DeepWiki entry points.
    • Updated OpenAgent project links, contact information, and acknowledgments.
    • Continued aligning the repository around the OpenAgent product identity, logo, and public-facing documentation.

    Fixed

    • Fixed stale home assistant intent cache behavior by improving message-signature based checks.
    • Fixed unstable home message history pagination caused by SQLAlchemy Row / tuple ID compatibility edge cases.
    • Fixed home navigation behavior when starting a new conversation from the sidebar.
    • Fixed model metadata display inconsistencies.
    • Fixed public app and public workflow update stability issues.
    • Reduced websocket transport race conditions during notification connection set
  5. v1.1.02026-04-12 · 2026-04-12 17:32 UTC
    查看更新说明

    OpenAgent v1.1.0

    This release focuses on public-agent orchestration, natural-language AI app creation, security hardening, deployment reliability, and a complete README/documentation refresh.

    Highlights

    • Home Assistant is now a stronger AI entry point for two core workflows:
      • Route user questions to the most relevant published public agents through A2A
      • Turn natural-language requirements into new AI app / agent creation flows
    • Improved public agent routing, indexing, and delegation flow for published apps
    • Expanded OpenAgent product presentation with bilingual README updates, architecture diagram, product overview GIF, and clearer OpenAPI/deployment guidance

    Features

    • Added A2A-oriented assistant agent flow and public agent delegation support
    • Added public agent registry and retrieval/index synchronization flow
    • Added agent notification support for auto-created apps
    • Added natural-language app creation flow backed by asynchronous task execution
    • Added email registration flow and related auth improvements
    • Added improved conversation search and Home Assistant interaction flows
    • Added richer app / workflow / store / OpenAPI frontend experience

    Fixes

    • Hardened email authentication, registration, and login flows
    • Prevented auth login enumeration and restored environment example coverage
    • Unified API and WebSocket endpoint resolution
    • Fixed Docker UI build argument handling and related environment propagation
    • Restored Docker environment handling and WebSocket dependency setup
    • Merged Alembic heads and added upgrade verification
    • Removed personal information from .env.example

    Documentation

    • Reworked README.md and README_ZH.md
    • Added a product overview GIF and architecture section
    • Highlighted A2A routing and natural-language app creation in the Home Assistant section
    • Improved Docker and environment setup documentation

    Notes

    • This release is intended as the new stable release after v1.0.0
    • If you are self-hosting, review api/.env.example and docker/README.md before deployment

    Full Changelog

    Compare: v1.0.0...v1.1.0

用户点评

登录用户的真实使用记录，可按版本和硬件筛选
共 0 条

    还没有人点评。第一个说出真实体验的人最有价值。

    README 原文

    未翻译
    在 GitHub 打开

    Table of Contents

    About The Project

    OpenAgent is a full-stack platform for teams building AI applications rather than a single chat demo. The repository combines a Flask backend, Celery workers, a Vue 3 frontend, visual workflow authoring, dataset and document management, public app and workflow publishing, and OpenAPI-based delivery.

    What the current codebase already supports:

    • Use the home assistant to route user requests to published public agents through A2A, or turn natural-language requirements into new AI app creation flows.
    • Build and manage AI apps from a dedicated workspace with draft, publish, analysis, version comparison, and prompt comparison flows.
    • Enable Deep Research to let the app decompose complex tasks and coordinate bound capabilities across multi-step execution, suitable for scenarios requiring deep reasoning and concrete outputs.
    • Design workflows visually with nodes for LLMs, tool calls, dataset retrieval, code execution, HTTP requests, branching, text processing, template transforms, and structured parameter extraction.
    • Manage datasets, upload documents, inspect segments, and connect retrieval to agents and workflows.
    • Browse public apps, tools, and workflows through store-style views.
    • Expose published apps over REST and SSE through POST /api/openapi/chat.

    Architecture

    Click the diagram to view the full-resolution architecture image.

    See the Documentation index for source-level guides and the Chinese documentation index.

    Detailed source-level guide: Architecture documentation.

    Built With

    • AI framework and orchestration: LangChain, LangGraph, workflow orchestration, tool calling, A2A delegation, skills, memory
    • Knowledge and retrieval: RAG, semantic retrieval, full-text retrieval, hybrid retrieval, Weaviate, FAISS
    • Backend: Python, Flask, SQLAlchemy, Celery, Flask-SocketIO, Redis, PostgreSQL
    • Frontend: Vue 3, JavaScript / TypeScript, Vite, TailwindCSS, Pinia, Vue Flow, Arco Design
    • Infrastructure and delivery: Docker Compose, Nginx, OpenAPI, SSE
    • Model integrations: OpenAI, Atlas Cloud, DeepSeek, Grok, Google, Moonshot, Tongyi, Wenxin, Ollama, Zhipu

    Sponsored By

    Atlas Cloud
    Bloome
    • Bloome is an AI-agent IM platform where you can interact with AI agents and build collaborative communities. Official website: Bloome

    Getting Started

    Prerequisites

    • Docker 20.10+
    • Docker Compose 2.x
    • 8 GB+ RAM recommended for the full stack
    • Access to at least one supported model provider API key

    Installation

    1. Clone the repository.

      git clone https://github.com/Haohao-end/openagent.git
cd openagent

    2. Create the runtime environment file.

      cp api/.env.example api/.env

    3. Review the minimum required settings in api/.env.

      • JWT_SECRET_KEY
      • POSTGRES_PASSWORD
      • REDIS_PASSWORD
      • WEAVIATE_API_KEY
      • VITE_API_PREFIX
      • At least one provider key such as OPENAI_API_KEY, ATLASCLOUD_API_KEY, DEEPSEEK_API_KEY, or DASHSCOPE_API_KEY

    4. Start the Docker stack.

      cd docker
docker compose up -d --build

    5. Open the local services.

      ServiceURLNotes
      Frontendhttp://localhost:3000Vue 3 web UI
      APIhttp://localhost:5001Flask REST API
      Nginxhttp://localhostReverse proxy

    Local Development

    For local development, start only the supporting Docker services; the API and UI run directly on your machine. This avoids building or starting the API, UI, Celery, or Nginx containers:

    cd docker
docker compose up -d --no-build llmops-db llmops-redis llmops-weaviate
docker compose ps llmops-db llmops-redis llmops-weaviate

    Use localhost and the published ports for PostgreSQL, Redis, and Weaviate in your local api/.env configuration. See Docker setup for details. Keep full-stack docker compose up -d --build for containerized deployment, not this local workflow.

    Backend:

    cd api
pip install -r requirements.txt
flask run --port 5001

    Frontend:

    cd ui
npm install
npm run serve

    Vite serves the frontend on port 5173 by default. The frontend configuration resolves the API base from VITE_API_PREFIX, and local development commonly proxies /api to the Flask backend.

    Useful commands:

    cd api
pytest

    cd ui
npm run type-check
npm run lint
npm run build
npm run test:unit -- --run

    Usage

    1. Home Assistant Experience

    Use the home page as the default assistant entry point to route user questions to the most relevant published public agents through A2A, or describe a new idea in natural language and trigger AI app creation. The same surface also supports multi-turn chat, suggested prompts, image upload, and audio input.

    2. App Workspace and Deep Research

    The app workspace is the main work area for an AI app, not a standalone settings page. The left side handles model, prompt, and capability bindings. The right side is used for live debugging, execution traces, and result checks. In the current codebase, the README term Deep Research maps to the deep thinking mode behind enable_deep_thinking.

    Key capabilities:

    • Configuration and version management: manage model changes, prompt logic, drafts, publishing, version comparison, prompt comparison, and app duplication in one place.
    • Capability integration: bind plugins, MCP, Skills, child Agent apps, workflows, and datasets through a single workspace.
    • Complex task execution: with Deep Research enabled, the app can break work into steps and coordinate bound capabilities across a longer execution chain.
    • Sandbox and artifact output: support script execution, code handling, file generation, and attachment export for tasks that need concrete outputs.
    • Debugging and result verification: use the right-side panel to run real conversations and inspect deep execution timelines, task state, produced artifacts, and final results.

    3. Visual Workflow Editor

    Author workflows with nodes such as LLM, tool, dataset retrieval, code, HTTP request, template transform, text processor, variable assigner, parameter extractor, if/else, start, and end.

    4. Dataset and Retrieval

    Create datasets, upload documents, inspect segments, and wire retrieval nodes into workflows or AI apps for knowledge-enabled behavior.

    5. OpenAPI Delivery

    Publish an app and call it over POST /api/openapi/chat with standard or streaming responses, including support for multi-turn conversation identifiers.

    Testing

    The repository already includes automated backend and frontend tests.

    • Backend: cd api && pytest
    • Frontend unit tests: cd ui && npm run test:unit -- --run
    • Frontend type check: cd ui && npm run type-check
    • Frontend build validation: cd ui && npm run build

    Contact

    Acknowledgments

    • Thanks to Atlas Cloud and Bloome for supporting OpenAgent.
    • Special thanks to Rui Yang and Haoyu Wang (Johns Hopkins University) for responsibly reporting a Host Header poisoning issue in the built-in tool icon URL construction and helping improve the security of this project.
    • Special thanks to Ryan B for responsibly reporting an SSRF vulnerability in the custom API-tool server URL and helping improve the project's outbound request security.
    • Thanks to 汪钶楠 (Wang Kenan) for reporting an AppleScript argument-handling concern in the macOS screenshot helper and helping us verify a safe fix.

    相关项目

    同分类里 star 最高的项目
    对比全部 →
    langfuse
    langfuse·TypeScript

    开源 LLM 应用可观测与评估平台，能追踪调用链路、跑评测实验，帮团队定位问题并持续优化提示词和模型效果。

    LLM 可观测+4LLM 可观测
    37261+1374.2kv4.77.02026-11-10 17:13 UTC
    signoz
    SigNoz·TypeScript

    基于 OpenTelemetry 的开源可观测性平台，把日志、指标和链路追踪集中到一处，提供 APM、分布式追踪和基础设施监控，帮团队快速定位线上问题。

    可观测性+4LLM 可观测
    32570+262.6kv0.149.02026-11-10 16:52 UTC
    skywalking
    apache·Java

    分布式系统的应用性能监控平台，支持链路追踪、指标采集和日志分析，帮你在微服务架构下快速定位性能瓶颈。

    应用性能监控+4LLM 可观测
    24999+56.6kv11.1.02026-11-10 12:30 UTC
    jaeger
    jaegertracing·Go

    CNCF 出品的分布式追踪平台，采集服务调用链路并做可视化分析，帮你在微服务和 LLM 应用里定位延迟与故障根因。

    分布式追踪+4LLM 可观测
    23394+83.2kv2.23.02026-11-10 17:10 UTC