[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-9060":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":16,"stars7d":16,"stars30d":17,"stars90d":16,"forks30d":16,"starsTrendScore":16,"compositeScore":18,"rankGlobal":10,"rankLanguage":10,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":20,"hasPages":20,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":27,"readmeContent":28,"aiSummary":29,"trendingCount":16,"starSnapshotCount":16,"syncStatus":30,"lastSyncTime":31,"discoverSource":32},9060,"vector-vein","AndersonBY\u002Fvector-vein","AndersonBY","No-code AI workflow. Drag and drop workflow nodes and use your workflow with your AI agents.","https:\u002F\u002Fvectorvein.ai",null,"Vue",942,143,9,8,0,1,50.58,"Other",false,"main",[23,24,25,26],"ai","openai","python","workflow","2026-06-12 04:00:42","English | [简体中文](README_zh.md) | [日本語](README_ja.md)\n\n[![VectorVein](resources\u002Fimages\u002Fvector-vein-with-text-primary-en.svg)](https:\u002F\u002Fvectorvein.ai)\n\n# 🔀 VectorVein\n\nBuild your automation workflow with the power of AI and your personal knowledge base.\n\nCreate powerful workflows with just drag and drop, without any programming.\n\n[![Online Version of VectorVein](resources\u002Fimages\u002Fdemo-en.gif)](https:\u002F\u002Fgithub.com\u002FAndersonBY\u002Fvector-vein)\n\nVectorVein is a no-code AI workflow software inspired by [LangChain](https:\u002F\u002Fgithub.com\u002Fhwchase17\u002Flangchain) and [langflow](https:\u002F\u002Fgithub.com\u002Flogspace-ai\u002Flangflow), designed to combine the powerful capabilities of large language models and enable users to easily achieve intelligent and automated workflows for various daily tasks.\n\n## 🌐 Online Experience\n\nYou can experience VectorVein's online version [here](https:\u002F\u002Fvectorvein.ai\u002Fen), with no need to download or install.\n\nOfficial website [Online Documentation](https:\u002F\u002Fvectorvein.ai\u002Fhelp\u002Fdocs\u002Fintroduction)\n\n## 📦 Installation and Configuration\n\n### Installation\n\nAfter downloading VectorVein from [Release](https:\u002F\u002Fgithub.com\u002FAndersonBY\u002Fvector-vein\u002Freleases\u002F), the program will create a \"data\" folder in the installation directory to store the database and static file resources.\n\nVectorVein is built using pywebview, based on the webview2 kernel, so you need to install the webview2 runtime. If the software cannot be opened, you may need to download the webview2 runtime manually from [https:\u002F\u002Fdeveloper.microsoft.com\u002Fen-us\u002Fmicrosoft-edge\u002Fwebview2\u002F](https:\u002F\u002Fdeveloper.microsoft.com\u002Fen-us\u002Fmicrosoft-edge\u002Fwebview2\u002F)\n\n> [!IMPORTANT]\n> If the software cannot be opened after decompression, please check if the downloaded compressed package .zip file is locked. You can solve this problem by right-clicking the compressed package and selecting \"Unblock\".\n\n### Configuration\n\nMost workflows and agents in the software involve the use of AI large language models, so you should at least provide a usable configuration for a large language model. For workflows, you can see which large language models are being used in the interface, as shown in the image below.\n\n![LLM used in workflow](resources\u002Fimages\u002Fworkflow-llm-use-en.jpg)\n\n#### API Endpoint Configuration\n\nStarting from v0.2.10, VectorVein separates API endpoints and large language model configurations, allowing multiple API endpoints for the same large language model.\n\n![API Endpoint Configuration](resources\u002Fimages\u002Fendpoint-settings_en-US.jpg)\n\nAfter the software opens normally, click the open settings button, and you can configure the information for each API endpoint as needed, or add custom API endpoints. Currently, the API endpoints support OpenAI-compatible interfaces, which can be connected to locally running services such as LM-Studio, Ollama, vLLM, etc.\n\n> The API Base for LM-Studio is typically http:\u002F\u002Flocalhost:1234\u002Fv1\u002F\n> \n> The API Base for Ollama is typically http:\u002F\u002Flocalhost:11434\u002Fv1\u002F\n\n#### Remote Large Language Model Interface Configuration\n\nPlease configure the specific information for each model in the `Remote LLMs` tab.\n\n![LLM Settings](resources\u002Fimages\u002Fremote-llms-settings_en-US.jpg)\n\nClick on any model to set its specific configuration, as shown below.\n\n![LLM Settings](resources\u002Fimages\u002Fremote-llms-settings-2_en-US.jpg)\n\n> The `Model Key` is the standard name of the large model and generally does not need to be adjusted. The `Model ID` is the name used during actual deployment, which usually matches the `Model Key`. However, in deployments like Azure OpenAI, the `Model ID` is user-defined and therefore needs to be adjusted according to the actual situation.\n>\n> Since the model IDs from different providers for the same model may vary, you can click the `Edit` button to configure the specific model ID under this endpoint, as shown in the figure below.\n>\n> ![Endpoint Model ID Configuration](resources\u002Fimages\u002Fendpoint-model-id-settings_en-US.jpg)\n\n#### Custom Large Language Model Interface Configuration\n\nIf using a custom large language model, fill in the custom model configuration information on the `Custom LLMs` tab. Currently, interfaces compatible with OpenAI are supported, such as LM-Studio, Ollama, vLLM, etc.\n\n![Custom LLM Settings](resources\u002Fimages\u002Fcustom-llms-settings_en-US.jpg)\n\nFirst, add a custom model family, then add a custom model. Don't forget to click the `Save Settings` button.\n\n#### Speech Recognition Configuration\n\nCurrently, speech recognition uses the OpenAI-compatible path. You can use the same configuration as the large language model or set up a speech recognition service compatible with the OpenAI API (such as Groq).\n\n![Speech Recognition Configuration](resources\u002Fimages\u002Fasr-settings1-en.jpg)\n\n### Embedding Configuration\n\nWhen you need to perform vector searches using vector data, configure embedding backends in the `Embedding models` settings through the `vv-llm` `embedding_backends` scheme. Built-in OpenAI embeddings and custom request\u002Fresponse mappings are supported, so local services such as [text-embeddings-inference](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftext-embeddings-inference) can be connected by configuring endpoints plus request\u002Fresponse mappings.\n\n![Local Embedding Settings](resources\u002Fimages\u002Fembedding-settings1-en.jpg)\n\n### Shortcut Settings\n\nFor ease of daily use, you can configure shortcuts to quickly initiate voice conversations with the Agent. By launching through the shortcut, you can directly interact with the Agent via speech recognition. It is important to ensure that the speech recognition service is correctly configured beforehand.\n\n**Include Screenshot** means that while starting the conversation, a screenshot of the screen will be taken and uploaded as an attachment to the conversation.\n\n![Shortcut Settings](resources\u002Fimages\u002Fshortcut-settings1-en.jpg)\n\n### Notes\n\n#### About the local Stable Diffusion API\n\nTo use your own local Stable Diffusion API, you need to add the parameter --api to the startup item of webui-user.bat, that is\n\n```\nset COMMANDLINE_ARGS=--api\n```\n\n## 💻 Usage\n\n### 🔌 API Access (New in v0.4.0)\n\nVectorVein now provides a local API service that allows you to call workflows programmatically. This enables integration with other applications and automation tools.\n\n#### API Features\n\n- **Local FastAPI Server**: Runs automatically when VectorVein starts\n- **RESTful Interface**: Standard HTTP endpoints for workflow operations\n- **Workflow Execution**: Run workflows with custom input parameters\n- **Status Monitoring**: Check workflow execution status and results\n- **OpenAPI Documentation**: Interactive API documentation at `\u002Fdocs`\n\n#### API Endpoints\n\nThe API service runs on `http:\u002F\u002Flocalhost:8787` (default port) and provides the following endpoints:\n\n- `GET \u002Fapi\u002Finfo` - Get API server information\n- `GET \u002Fapi\u002Fworkflow\u002Flist` - List all workflows\n- `GET \u002Fapi\u002Fworkflow\u002F{workflow_id}` - Get workflow details\n- `POST \u002Fapi\u002Fworkflow\u002Frun` - Execute a workflow\n- `POST \u002Fapi\u002Fworkflow\u002Fcheck-status` - Check workflow execution status\n- `GET \u002Fhealth` - Health check endpoint\n\n#### Example Usage\n\n```python\nimport requests\n\n# Run a workflow\nresponse = requests.post('http:\u002F\u002Flocalhost:8787\u002Fapi\u002Fworkflow\u002Frun', json={\n    'wid': 'your-workflow-id',\n    'input_fields': [\n        {'node_id': 'node1', 'field_name': 'input', 'value': 'Hello World'}\n    ],\n    'wait_for_completion': True\n})\n\nresult = response.json()\nprint(result['data'])  # Workflow output\n```\n\nFor detailed API documentation, visit `http:\u002F\u002Flocalhost:8787\u002Fdocs` after starting VectorVein.\n\n### 📖 Basic Concepts\n\nA workflow represents a work task process, including input, output, and how input is processed to reach the output result.\n\nExamples:\n\n- **Translation Workflow**: The input is an English Word document, and the output is also a Word document. You can design a workflow to translate the input Chinese document and generate a Chinese document output.\n- **Mind Map Workflow**: If the output of the translation workflow is changed to a mind map, you can get a workflow that reads an English Word document and summarizes it into a Chinese mind map.\n- **Web Article Summary Workflow**: If the input of the mind map workflow is changed to a URL of a web article, you can get a workflow that reads a web article and summarizes it into a Chinese mind map.\n- **Automatic Classification of Customer Complaints Workflow**: The input is a table containing complaint content, and you can customize the keywords that need to be classified, so that the complaints can be automatically classified. The output is an automatically generated Excel table containing the classification results.\n\n### 🔎 User Interface\n\nEach workflow has a **User Interface** and an **Editor Interface**. The user interface is used for daily workflow operations, and the editor interface is used for workflow editing. Usually, after designing a workflow, you only need to run it in the user interface and do not need to modify it in the editor interface.\n\n![User Interface](resources\u002Fimages\u002Fuser-interface1-en.jpg)\n\nThe user interface is shown above and is divided into three parts: input, output, and trigger (usually a run button). You can directly enter content for daily use, click the run button to see the output result.\n\nTo view the executed workflow, click **Workflow Run Records**, as shown in the following figure.\n\n![Workflow Run Records](resources\u002Fimages\u002Fworkflow-record-en.jpg)\n\n### ✏️ Creating a Workflow\n\nYou can add our official templates to your workflow or create a new one. It is recommended to familiarize yourself with the use of workflows using official templates at the beginning.\n\n![Workflow Editor Interface](resources\u002Fimages\u002Feditor-en.jpg)\n\nThe workflow editor interface is shown above. You can edit the name, tags, and detailed description at the top. The left side is the node list of the workflow, and the right is the canvas of the workflow. You can drag the desired node from the left side to the canvas, and then connect the node through the wire to form a workflow.\n\nYou can view a tutorial on creating a simple crawler + AI summary mind map workflow [here](TUTORIAL.md).\n\nYou can also try this [online interactive tutorial](https:\u002F\u002Fvectorvein.ai\u002Fworkspace\u002Fworkflow\u002Feditor\u002Ftutorial).\n\n## 🛠️ Development and Deployment\n\n### Environment Requirements\n\n- Backend\n  - Python 3.8 ~ Python 3.11\n  - [PDM](https:\u002F\u002Fpdm.fming.dev\u002Flatest\u002F#installation) installed\n\n- Frontend\n  - Vue3\n  - Vite\n\n### Project Development\n\nCopy and modify backend\u002F.env.example to .env file, this is the basic environment variable information, which will be used during development and packaging.\n\nRun the following command in the **backend** directory to install dependencies:\n\n#### Windows\n```bash\npdm install\n```\n\n#### Mac\n```bash\npdm install -G mac\n```\n\nNormally, PDM will automatically find the system's Python and create a virtual environment and install dependencies.\n\nAfter installation, run the following command to start the backend development server and see the running effect:\n\n```bash\npdm run dev\n```\n\nFor backend quality checks, you can run the following commands in the `backend` directory:\n\n```bash\npdm run test\npdm run lint\npdm run typecheck\n```\n\nIf you need to modify the frontend code, you need to run the following command in the **frontend** directory to install dependencies:\n\n```bash\npnpm install\n```\n\nBefore submitting frontend changes, run the following quality checks in the **frontend** directory:\n\n```bash\npnpm run lint\npnpm exec vite build\n```\n\n> When pulling the project code for the first time, you also need to run `pnpm install` to install the front-end dependencies.\n>\n> If you don't need to develop any front-end code at all, you can directly copy the `web` folder from the release version into the `backend` folder.\n\nAfter the frontend dependencies are installed, you need to compile the frontend code into the static file directory of the backend. A shortcut instruction has been provided in the project. Run the following command in the **backend** directory to pack and copy the frontend resources:\n\n```bash\npdm run build-front\n```\n\n### Database Structure Changes\n\n> [!WARNING]\n> Before making changes to the database structure, please back up your database (located at `my_database.db` in your configured `data` directory), otherwise you may lose data.\n\nIf you have modified the model structure in `backend\u002Fmodels`, you need to run the following commands in the `backend` directory to update the database structure:\n\nFirst, enter the Python environment:\n\n```bash\npdm run python\n```\n\n```python\nfrom models import create_migrations\ncreate_migrations(\"migration_name\")  # Name according to the changes made\n```\n\nAfter the operation, a new migration file will be generated in the `backend\u002Fmigrations` directory, with the filename format `xxx_migration_name.py`. It is recommended to check the content of the migration file first to ensure it is correct, and then restart the main program. The main program will automatically execute the migration.\n\n### Software Packaging\n\nThe project uses pyinstaller for packaging. Run the following command in the **backend** directory to package it into an executable file:\n\n```bash\npdm run build\n```\n\nAfter packaging, the executable file will be generated in the**backend\u002Fdist** directory.\n\n### Automated GitHub Releases\n\nThe repository now includes a GitHub Actions workflow at `.github\u002Fworkflows\u002Frelease.yml` for tagged releases.\n\n1. Update `backend\u002Fpyproject.toml` so `project.version` matches the release version you want to publish.\n2. Commit the version change.\n3. Create and push a tag like `v0.4.3`.\n4. GitHub Actions will automatically build Windows, macOS, and Linux ZIP packages and attach them to the matching GitHub Release.\n\nIf the tag version and `backend\u002Fpyproject.toml` version do not match, the workflow will fail fast before packaging.\n\n## 📄 License\n\nVectorVein is an open-source software that supports personal non-commercial use. Please refer to [LICENSE](LICENSE.md) for specific agreements.\n","VectorVein 是一个无代码的AI工作流构建工具，允许用户通过拖拽节点来创建和使用AI代理的工作流。它基于Vue开发，支持与大型语言模型集成，使非编程人员也能轻松实现日常任务的自动化处理。该软件提供了一个直观的图形界面，让用户无需编写代码即可设计出复杂的业务流程，并且支持多样的API端点配置，方便接入不同的AI服务。适用于需要快速搭建基于AI的自动化解决方案但又缺乏专业编程技能的个人或团队。",2,"2026-06-11 03:21:00","top_language"]