[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tool-epsilla-cloud--vectordb":3,"similar-epsilla-cloud--vectordb":162},{"id":4,"github_repo":5,"name":6,"description_en":7,"description_zh":8,"ai_summary_zh":8,"readme_en":9,"readme_zh":10,"quickstart_zh":11,"use_case_zh":12,"hero_image_url":13,"owner_login":14,"owner_name":15,"owner_avatar_url":16,"owner_bio":17,"owner_company":18,"owner_location":18,"owner_email":18,"owner_twitter":19,"owner_website":20,"owner_url":21,"languages":22,"stars":43,"forks":44,"last_commit_at":45,"license":46,"difficulty_score":47,"env_os":48,"env_gpu":49,"env_ram":49,"env_deps":50,"category_tags":52,"github_topics":59,"view_count":47,"oss_zip_url":18,"oss_zip_packed_at":18,"status":77,"created_at":78,"updated_at":79,"faqs":80,"releases":86},1144,"epsilla-cloud\u002Fvectordb","vectordb","Epsilla is a high performance Vector Database Management System","Vectordb是一款面向人工智能应用的开源向量数据库系统，专注于提升向量数据的存储与检索效率。它通过优化算法和架构设计，解决了传统数据库在处理高维向量相似性搜索时的性能瓶颈，尤其适合需要实时语义理解的场景。开发者可以通过Python、JavaScript等语言直接调用，快速构建包含文本、图像等多模态数据的智能应用。系统采用云原生设计，支持混合搜索模式，能同时处理密集向量与稀疏特征，且提供与LangChain、LlamaIndex等框架的集成方案。其底层基于C++实现，利用并行图遍历技术，在保持99.9%搜索精度的同时，使向量检索速度较同类方案提升10倍以上。研究机构和企业开发者可借助它降低AI模型部署成本，而普通用户则能通过集成该系统的应用获得更精准的智能服务体验。","\u003Cdiv align=\"center\">\n\u003Cp align=\"center\">\n\n\u003Cimg width=\"275\" alt=\"Epsilla Logo\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fepsilla-cloud_vectordb_readme_6da23f048fe3.png\">\n\n**A 10x faster, cheaper, and better vector database**\n\n\u003Ca href=\"https:\u002F\u002Fepsilla-inc.gitbook.io\u002Fepsilladb\u002F\">Documentation\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FcDaY2CxZc5\">Discord\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Ftwitter.com\u002Fepsilla_inc\">Twitter\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fblog.epsilla.com\">Blog\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fwww.youtube.com\u002F@Epsilla-kp5cx\">YouTube\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fforms.gle\u002Fz73ra1sGBxH9wiUR8\">Feedback\u003C\u002Fa>\n\n\u003C\u002Fdiv>\n\n\u003Chr \u002F>\n\nEpsilla is an open-source vector database. Our focus is on ensuring scalability, high performance, and cost-effectiveness of vector search. EpsillaDB bridges the gap between information retrieval and memory retention in Large Language Models.\n\n## Quick Start using Docker\n\n**1. Run Backend in Docker**\n```shell\ndocker pull epsilla\u002Fvectordb\ndocker run --pull=always -d -p 8888:8888 -v \u002Fdata:\u002Fdata epsilla\u002Fvectordb\n```\n\n**2. Interact with Python Client**\n```shell\npip install pyepsilla\n```\n\n```python\nfrom pyepsilla import vectordb\n\nclient = vectordb.Client(host='localhost', port='8888')\nclient.load_db(db_name=\"MyDB\", db_path=\"\u002Fdata\u002Fepsilla\")\nclient.use_db(db_name=\"MyDB\")\n\nclient.create_table(\n    table_name=\"MyTable\",\n    table_fields=[\n        {\"name\": \"ID\", \"dataType\": \"INT\", \"primaryKey\": True},\n        {\"name\": \"Doc\", \"dataType\": \"STRING\"},\n    ],\n    indices=[\n      {\"name\": \"Index\", \"field\": \"Doc\"},\n    ]\n)\n\nclient.insert(\n    table_name=\"MyTable\",\n    records=[\n        {\"ID\": 1, \"Doc\": \"Jupiter is the largest planet in our solar system.\"},\n        {\"ID\": 2, \"Doc\": \"Cheetahs are the fastest land animals, reaching speeds over 60 mph.\"},\n        {\"ID\": 3, \"Doc\": \"Vincent van Gogh painted the famous work \\\"Starry Night.\\\"\"},\n        {\"ID\": 4, \"Doc\": \"The Amazon River is the longest river in the world.\"},\n        {\"ID\": 5, \"Doc\": \"The Moon completes one orbit around Earth every 27 days.\"},\n    ],\n)\n\nclient.query(\n    table_name=\"MyTable\",\n    query_text=\"Celestial bodies and their characteristics\",\n    limit=2\n)\n\n# Result\n# {\n#     'message': 'Query search successfully.',\n#     'result': [\n#         {'Doc': 'Jupiter is the largest planet in our solar system.', 'ID': 1},\n#         {'Doc': 'The Moon completes one orbit around Earth every 27 days.', 'ID': 5}\n#     ],\n#     'statusCode': 200\n# }\n```\n\n\n## Features:\n\n* High performance and production-scale similarity search for embedding vectors.\n\n* Full fledged database management system with familiar database, table, and field concepts. Vector is just another field type.\n\n* Metadata filtering.\n\n* Hybrid search with a fusion of dense and sparse vectors.\n\n* Built-in embedding support, with natural language in natural language out search experience.\n\n* Cloud native architecture with compute storage separation, serverless, and multi-tenancy.\n\n* Rich ecosystem integrations including LangChain and LlamaIndex.\n\n* Python\u002FJavaScript\u002FRuby clients, and REST API interface.\n\nEpsilla's core is written in C++ and leverages the advanced academic parallel graph traversal techniques for vector indexing, achieving 10 times faster vector search than HNSW while maintaining precision levels of over 99.9%.\n\n\n\n## Epsilla Cloud\n\nTry our fully managed vector DBaaS at \u003Ca href=\"https:\u002F\u002Fcloud.epsilla.com\u002F\">Epsilla Cloud\u003C\u002Fa>\n\n## (Experimental) Use Epsilla as a python library without starting a docker image\n\n**1. Build Epsilla Python Bindings lib package**\n```shell\ncd engine\u002Fscripts\n(If on Ubuntu, run this first: bash setup-dev.sh)\nbash install_oatpp_modules.sh\ncd ..\nbash build.sh\nls -lh build\u002F*.so\n```\n\n**2. Run test with python bindings lib \"epsilla.so\" \"libvectordb_dylib.so in the folder \"build\" built in the previous step**\n```shell\ncd engine\nexport PYTHONPATH=.\u002Fbuild\u002F\nexport DB_PATH=\u002Ftmp\u002Fdb33\npython3 test\u002Fbindings\u002Fpython\u002Ftest.py\n```\n\nHere are some sample code:\n```python\nimport epsilla\n\nepsilla.load_db(db_name=\"db\", db_path=\"\u002Fdata\u002Fepsilla\")\nepsilla.use_db(db_name=\"db\")\nepsilla.create_table(\n    table_name=\"MyTable\",\n    table_fields=[\n        {\"name\": \"ID\", \"dataType\": \"INT\", \"primaryKey\": True},\n        {\"name\": \"Doc\", \"dataType\": \"STRING\"},\n        {\"name\": \"EmbeddingEuclidean\", \"dataType\": \"VECTOR_FLOAT\", \"dimensions\": 4, \"metricType\": \"EUCLIDEAN\"}\n    ]\n)\nepsilla.insert(\n    table_name=\"MyTable\",\n    records=[\n        {\"ID\": 1, \"Doc\": \"Berlin\", \"EmbeddingEuclidean\": [0.05, 0.61, 0.76, 0.74]},\n        {\"ID\": 2, \"Doc\": \"London\", \"EmbeddingEuclidean\": [0.19, 0.81, 0.75, 0.11]},\n        {\"ID\": 3, \"Doc\": \"Moscow\", \"EmbeddingEuclidean\": [0.36, 0.55, 0.47, 0.94]}\n    ]\n)\n(code, response) = epsilla.query(\n    table_name=\"MyTable\",\n    query_field=\"EmbeddingEuclidean\",\n    response_fields=[\"ID\", \"Doc\", \"EmbeddingEuclidean\"],\n    query_vector=[0.35, 0.55, 0.47, 0.94],\n    filter=\"ID \u003C 6\",\n    limit=10,\n    with_distance=True\n)\nprint(code, response)\n```\n\n\n","\u003Cdiv align=\"center\">\n\u003Cp align=\"center\">\n\n\u003Cimg width=\"275\" alt=\"Epsilla Logo\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fepsilla-cloud_vectordb_readme_6da23f048fe3.png\">\n\n**一款速度提升10倍、成本更低、性能更优的向量数据库**\n\n\u003Ca href=\"https:\u002F\u002Fepsilla-inc.gitbook.io\u002Fepsilladb\u002F\">文档\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FcDaY2CxZc5\">Discord\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Ftwitter.com\u002Fepsilla_inc\">Twitter\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fblog.epsilla.com\">博客\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fwww.youtube.com\u002F@Epsilla-kp5cx\">YouTube\u003C\u002Fa> •\n\u003Ca href=\"https:\u002F\u002Fforms.gle\u002Fz73ra1sGBxH9wiUR8\">反馈\u003C\u002Fa>\n\n\u003C\u002Fdiv>\n\n\u003Chr \u002F>\n\nEpsilla 是一款开源的向量数据库。我们的核心目标是确保向量搜索的可扩展性、高性能和高性价比。EpsillaDB 桥接了大型语言模型中信息检索与记忆保持之间的鸿沟。\n\n## 使用 Docker 快速入门\n\n**1. 在 Docker 中运行后端**\n```shell\ndocker pull epsilla\u002Fvectordb\ndocker run --pull=always -d -p 8888:8888 -v \u002Fdata:\u002Fdata epsilla\u002Fvectordb\n```\n\n**2. 使用 Python 客户端进行交互**\n```shell\npip install pyepsilla\n```\n\n```python\nfrom pyepsilla import vectordb\n\nclient = vectordb.Client(host='localhost', port='8888')\nclient.load_db(db_name=\"MyDB\", db_path=\"\u002Fdata\u002Fepsilla\")\nclient.use_db(db_name=\"MyDB\")\n\nclient.create_table(\n    table_name=\"MyTable\",\n    table_fields=[\n        {\"name\": \"ID\", \"dataType\": \"INT\", \"primaryKey\": True},\n        {\"name\": \"Doc\", \"dataType\": \"STRING\"},\n    ],\n    indices=[\n      {\"name\": \"Index\", \"field\": \"Doc\"},\n    ]\n)\n\nclient.insert(\n    table_name=\"MyTable\",\n    records=[\n        {\"ID\": 1, \"Doc\": \"木星是我们太阳系中最大的行星。\"},\n        {\"ID\": 2, \"Doc\": \"猎豹是陆地上跑得最快的动物，速度可达每小时60英里以上。\"},\n        {\"ID\": 3, \"Doc\": \"文森特·梵高创作了著名的画作《星夜》。\"},\n        {\"ID\": 4, \"Doc\": \"亚马逊河是世界上最长的河流。\"},\n        {\"ID\": 5, \"Doc\": \"月球每27天绕地球公转一周。\"},\n    ],\n)\n\nclient.query(\n    table_name=\"MyTable\",\n    query_text=\"天体及其特征\",\n    limit=2\n)\n\n# 结果\n# {\n#     'message': '查询成功。',\n#     'result': [\n#         {'Doc': '木星是我们太阳系中最大的行星。', 'ID': 1},\n#         {'Doc': '月球每27天绕地球公转一周。', 'ID': 5}\n#     ],\n#     'statusCode': 200\n# }\n```\n\n\n## 核心特性：\n\n* 面向生产环境的高性能相似度搜索，专为嵌入向量设计。\n\n* 功能完备的数据库管理系统，采用熟悉的数据库、表和字段概念。向量只是另一种字段类型。\n\n* 元数据过滤功能。\n\n* 密集向量与稀疏向量融合的混合搜索。\n\n* 内置嵌入支持，实现自然语言输入、自然语言输出的搜索体验。\n\n* 原生云架构，计算与存储分离、无服务器模式以及多租户支持。\n\n* 丰富的生态集成，包括 LangChain 和 LlamaIndex。\n\n* 提供 Python、JavaScript 和 Ruby 客户端，以及 REST API 接口。\n\nEpsilla 的核心采用 C++ 编写，并利用先进的学术级并行图遍历技术进行向量索引，从而实现比 HNSW 快10倍的向量搜索速度，同时保持超过99.9%的精确度。\n\n\n\n## Epsilla Cloud\n\n立即试用我们完全托管的向量 DBaaS 服务：\u003Ca href=\"https:\u002F\u002Fcloud.epsilla.com\u002F\">Epsilla Cloud\u003C\u002Fa>\n\n## （实验性）无需启动 Docker 镜像即可直接使用 Epsilla 的 Python 库\n\n**1. 构建 Epsilla Python 绑定库包**\n```shell\ncd engine\u002Fscripts\n（如果在 Ubuntu 系统上，请先运行：bash setup-dev.sh）\nbash install_oatpp_modules.sh\ncd ..\nbash build.sh\nls -lh build\u002F*.so\n```\n\n**2. 使用上一步构建的“build”文件夹中的 Python 绑定库“epsilla.so”和“libvectordb_dylib.so”运行测试**\n```shell\ncd engine\nexport PYTHONPATH=.\u002Fbuild\u002F\nexport DB_PATH=\u002Ftmp\u002Fdb33\npython3 test\u002Fbindings\u002Fpython\u002Ftest.py\n```\n\n以下是一些示例代码：\n```python\nimport epsilla\n\nepsilla.load_db(db_name=\"db\", db_path=\"\u002Fdata\u002Fepsilla\")\nepsilla.use_db(db_name=\"db\")\nepsilla.create_table(\n    table_name=\"MyTable\",\n    table_fields=[\n        {\"name\": \"ID\", \"dataType\": \"INT\", \"primaryKey\": True},\n        {\"name\": \"Doc\", \"dataType\": \"STRING\"},\n        {\"name\": \"EmbeddingEuclidean\", \"dataType\": \"VECTOR_FLOAT\", \"dimensions\": 4, \"metricType\": \"EUCLIDEAN\"}\n    ]\n)\nepsilla.insert(\n    table_name=\"MyTable\",\n    records=[\n        {\"ID\": 1, \"Doc\": \"柏林\", \"EmbeddingEuclidean\": [0.05, 0.61, 0.76, 0.74]},\n        {\"ID\": 2, \"Doc\": \"伦敦\", \"EmbeddingEuclidean\": [0.19, 0.81, 0.75, 0.11]},\n        {\"ID\": 3, \"Doc\": \"莫斯科\", \"EmbeddingEuclidean\": [0.36, 0.55, 0.47, 0.94]}\n    ]\n)\n(code, response) = epsilla.query(\n    table_name=\"MyTable\",\n    query_field=\"EmbeddingEuclidean\",\n    response_fields=[\"ID\", \"Doc\", \"EmbeddingEuclidean\"],\n    query_vector=[0.35, 0.55, 0.47, 0.94],\n    filter=\"ID \u003C 6\",\n    limit=10,\n    with_distance=True\n)\nprint(code, response)\n```","# Epsilla Vector Database 快速上手指南\n\n## 环境准备\n- **系统要求**：支持 Docker 的系统（Linux\u002FmacOS\u002FWindows）\n- **前置依赖**：\n  - 已安装 Docker Engine\n  - Python 3.7+（用于客户端）\n  > 中国开发者建议：配置 Docker 国内镜像加速器（如阿里云），并使用清华源加速 Python 包安装\n\n## 安装步骤\n1. **运行 Epsilla 后端服务**（使用 Docker）：\n   ```shell\n   docker pull epsilla\u002Fvectordb\n   docker run --pull=always -d -p 8888:8888 -v \u002Fdata:\u002Fdata epsilla\u002Fvectordb\n   ```\n\n2. **安装 Python 客户端**（推荐国内镜像源）：\n   ```shell\n   pip install pyepsilla -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple\n   ```\n\n## 基本使用\n以下是最简示例，实现向量搜索：\n\n```python\nfrom pyepsilla import vectordb\n\nclient = vectordb.Client(host='localhost', port='8888')\nclient.load_db(db_name=\"MyDB\", db_path=\"\u002Fdata\u002Fepsilla\")\nclient.use_db(db_name=\"MyDB\")\n\nclient.create_table(\n    table_name=\"MyTable\",\n    table_fields=[\n        {\"name\": \"ID\", \"dataType\": \"INT\", \"primaryKey\": True},\n        {\"name\": \"Doc\", \"dataType\": \"STRING\"},\n    ],\n    indices=[\n      {\"name\": \"Index\", \"field\": \"Doc\"},\n    ]\n)\n\nclient.insert(\n    table_name=\"MyTable\",\n    records=[\n        {\"ID\": 1, \"Doc\": \"Jupiter is the largest planet in our solar system.\"},\n        {\"ID\": 2, \"Doc\": \"Cheetahs are the fastest land animals, reaching speeds over 60 mph.\"},\n        {\"ID\": 3, \"Doc\": \"Vincent van Gogh painted the famous work \\\"Starry Night.\\\"\"},\n        {\"ID\": 4, \"Doc\": \"The Amazon River is the longest river in the world.\"},\n        {\"ID\": 5, \"Doc\": \"The Moon completes one orbit around Earth every 27 days.\"},\n    ],\n)\n\nclient.query(\n    table_name=\"MyTable\",\n    query_text=\"Celestial bodies and their characteristics\",\n    limit=2\n)\n```","某电商公司正在开发智能客服系统，需实时从百万级产品文档中检索相关信息，以快速回答用户关于商品特性的查询。\n\n### 没有 vectordb 时\n- 搜索响应时间长达5秒以上，用户频繁超时中断对话，客服满意度下降25%。\n- 高峰期数据库负载激增，需额外采购3台服务器，月均成本增加1.2万元。\n- 搜索精度不足，常返回无关商品信息（如将\"手机电池\"误匹配到\"耳机续航\"），人工复核率高达40%。\n- 索引维护依赖手动调优，运维团队每周投入10小时优化HNSW参数。\n\n### 使用 vectordb 后\n- 搜索响应时间压缩至0.3秒内，用户等待时长减少94%，客服满意度提升至92%。\n- 单节点轻松承载万级并发请求，服务器成本降低45%，月省5500元。\n- 精度提升至99.9%，精准匹配商品特性（如\"电池续航\"直接关联手机参数），人工复核率降至8%。\n- Docker一键部署+自动索引优化，运维时间从每周10小时降至1小时。\n\nvectordb 以10倍性能优势和显著成本节约，让电商智能客服系统实现毫秒级精准响应。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fepsilla-cloud_vectordb_ae1e348c.png","epsilla-cloud","Epsilla","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fepsilla-cloud_13424aeb.png","Hippocampus For AI",null,"epsilla_inc","https:\u002F\u002Fepsilla.com","https:\u002F\u002Fgithub.com\u002Fepsilla-cloud",[23,27,31,35,39],{"name":24,"color":25,"percentage":26},"C++","#f34b7d",97.4,{"name":28,"color":29,"percentage":30},"Python","#3572A5",1,{"name":32,"color":33,"percentage":34},"Shell","#89e051",0.8,{"name":36,"color":37,"percentage":38},"CMake","#DA3434",0.7,{"name":40,"color":41,"percentage":42},"Dockerfile","#384d54",0.1,872,44,"2026-04-04T14:07:10","GPL-3.0",3,"Linux","未说明",{"notes":49,"python":49,"dependencies":51},[],[53,54,55,56,57,58],"语言模型","图像","开发框架","其他","Agent","数据工具",[60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76],"ai","infrastructure","llms","chatgpt","data","data-science","database","embeddings","machine-learning","rag","retrieval","vector-database","embeddings-similarity","neural-network","neural-search","search-engine","vector-search","ready","2026-03-27T02:49:30.150509","2026-04-06T05:36:35.484535",[81],{"id":82,"question_zh":83,"answer_zh":84,"source_url":85},5173,"如何在 Python 客户端释放数据库内存而不关闭数据库？","使用 unload_db 方法可以释放数据库内存但保留数据。维护者建议：'unload_db 会释放数据库内存但保留数据。' 例如，调用 client.unload_db(db_name='your_db_name')。若需在上下文管理器中使用，可参考以下代码：\n```python\nclass epsillavdb():\n  def __init__(self, host=\"localhost\", port=\"8888\", dbname=\"\", db_path=f\"\u002Fdata\u002Fepsilla\u002F\"):\n    self.conn = vectordb.Client(host=host, port=port)\n  def __enter__(self):\n    status_code, _ = self.conn.load_db(db_name=dbname, db_path=db_path)\n    return self\n  def __exit__(self, exc_type, exc_val, exc_tb):\n    self.conn.unload_db(db_name=dbname)\n```","https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fissues\u002F135",[87,92,97,102,107,112,117,122,127,132,137,142,147,152,157],{"id":88,"version":89,"summary_zh":90,"released_at":91},104692,"cc-0.3.16","## What's Changed\r\n* Support string includes by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F145\r\n* Pre-filter support by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F147\r\n* add tools for object storage by @eric-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F148\r\n* Update README.md by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F152\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.12...cc-0.3.16","2025-09-07T15:47:22",{"id":93,"version":94,"summary_zh":95,"released_at":96},104693,"v0.3.16","## What's Changed\r\n* Support string includes by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F145\r\n* Pre-filter support by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F147\r\n* add tools for object storage by @eric-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F148\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.12...v0.3.16","2025-03-09T00:32:48",{"id":98,"version":99,"summary_zh":100,"released_at":101},104694,"v0.3.12","## What's Changed\r\n* Geospatial search by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F139\r\n* LIKE operator support by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F140\r\n* Facet (group by \u002F aggregation) support by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F142\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.11...v0.3.12","2024-04-03T13:56:31",{"id":103,"version":104,"summary_zh":105,"released_at":106},104695,"v0.3.11","## What's Changed\r\n* Create build-debug.yml by @eric-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F131\r\n* handle single quote in string value expr; by @ricki-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F133\r\n* Release API by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F132\r\n* Use logger to print logs by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F137\r\n* API to dump a database by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F138\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.9...v0.3.11","2024-03-14T18:25:37",{"id":108,"version":109,"summary_zh":110,"released_at":111},104696,"v0.3.9","## What's Changed\r\n* API to tune DB parameters by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F130\r\n* Hybrid search API\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.8...v0.3.9","2024-02-15T15:37:31",{"id":113,"version":114,"summary_zh":115,"released_at":116},104697,"v0.3.8","## What's Changed\r\n* Support for Nomic Embedding API and Nomic Embed by @AndriyMulyar in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F125\r\n* Support MistralAI embedding by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F128\r\n* Support variable embedding dimensions by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F129\r\n* Filter on distance by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F126\r\n\r\n## New Contributors\r\n* @AndriyMulyar made their first contribution in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F125\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.7...v0.3.8","2024-02-13T14:28:08",{"id":118,"version":119,"summary_zh":120,"released_at":121},104698,"v0.3.7","## What's Changed\r\n* Protect catalog change for high throughput by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F121\r\n* feat: Add mixedbread ai support by @juliuslipp in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F122\r\n  * Support uae-large-v1, bge-large-en-v1.5, gte-large, e5-large-v2, multilingual-e5-large, multilingual-e5-base, and gte-large-zh models\r\n* Support OpenAI new text-embedding-ada-003 and small embedding models\r\n* Support VoyageAI new voyage-2 and voyage-2-code embedding models\r\n\r\n## New Contributors\r\n* @juliuslipp made their first contribution in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F122\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.5...v0.3.7","2024-01-28T19:37:55",{"id":123,"version":124,"summary_zh":125,"released_at":126},104699,"v0.3.5","## What's Changed\r\n* Support Voyage AI embeddings by @richard-epsilla","2024-01-10T16:23:17",{"id":128,"version":129,"summary_zh":130,"released_at":131},104700,"v0.3.4","## What's Changed\r\n* Support upsert operation by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F117\r\n* Sequentialize data updates with mutex by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F118\r\n* Release 0.3.4 by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F120\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.3...v0.3.4","2024-01-06T16:59:56",{"id":133,"version":134,"summary_zh":135,"released_at":136},104701,"v0.3.3","## What's Changed\r\n* Fix dense vector index: fixed a bug that dense vector fields defined after sparse vector fields will cause coredump\r\n* Improve json.getInt function to automatically convert floating numbers into integers\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.2...v0.3.3","2023-12-31T04:33:46",{"id":138,"version":139,"summary_zh":140,"released_at":141},104702,"v0.3.2","## What's Changed\r\n* Support Jina AI embedding by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F106\r\n* Statistics API by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F109\r\n* Improve insert\u002Fdelete API response by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F113\r\n* Fix distance error in concurrent query by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F115\r\n* Fix delete by filter bug with int key by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F114\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.3.1...v0.3.2","2023-12-25T04:03:45",{"id":143,"version":144,"summary_zh":145,"released_at":146},104703,"v0.3.1","## What's Changed\r\n* Fix primary key insertion issue by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F98\r\n* Update VERSION by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F99\r\n* Auto embedding by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F100\r\n* integrate embed into vectordb by @eric-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F101\r\n\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.2.3...v0.3.1","2023-12-18T14:59:04",{"id":148,"version":149,"summary_zh":150,"released_at":151},104704,"v0.2.3","## What's Changed\r\n* Fix primary key insertion issue by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F98\r\n* Update README.md by @jonherke in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F96\r\n* Update README.md by @richard-epsilla in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F97\r\n\r\n## New Contributors\r\n* @jonherke made their first contribution in https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fpull\u002F96\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fepsilla-cloud\u002Fvectordb\u002Fcompare\u002Fv0.2.2...v0.2.3","2023-12-13T03:25:08",{"id":153,"version":154,"summary_zh":155,"released_at":156},104705,"v0.2.2","Announcing Epsilla v0.2, a solid step towards hybrid search support.\r\n\r\n## What's Changed\r\n✅ Support sparse vector.\r\n✅ Support dot product distance.\r\n✅ Multiple improvements and bug fixes.","2023-12-08T02:19:38",{"id":158,"version":159,"summary_zh":160,"released_at":161},104706,"v0.1.0","Announcing Epsilla v0.1, our first production oriented release!\r\nThis marks the milestone of a scalable, fully distributed version of Epsilla.\r\n\r\n## What's Changed\r\n✅ Sharding vector data for efficient horizontal scaling.\r\n✅ Multi-replica configuration for high availability and increased throughput.\r\n✅ Dynamic sharding expansion and auto-scaling replicas to meet data growth and traffic demands.\r\n\r\n","2023-11-21T11:50:15",[163,171,180,188,196,207],{"id":164,"name":165,"github_repo":166,"description_zh":167,"stars":168,"difficulty_score":47,"last_commit_at":169,"category_tags":170,"status":77},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[55,54,57],{"id":172,"name":173,"github_repo":174,"description_zh":175,"stars":176,"difficulty_score":177,"last_commit_at":178,"category_tags":179,"status":77},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[55,57,53],{"id":181,"name":182,"github_repo":183,"description_zh":184,"stars":185,"difficulty_score":177,"last_commit_at":186,"category_tags":187,"status":77},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[55,54,57],{"id":189,"name":190,"github_repo":191,"description_zh":192,"stars":193,"difficulty_score":177,"last_commit_at":194,"category_tags":195,"status":77},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[55,53],{"id":197,"name":198,"github_repo":199,"description_zh":200,"stars":201,"difficulty_score":177,"last_commit_at":202,"category_tags":203,"status":77},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[54,58,204,205,57,56,53,55,206],"视频","插件","音频",{"id":208,"name":209,"github_repo":210,"description_zh":211,"stars":212,"difficulty_score":47,"last_commit_at":213,"category_tags":214,"status":77},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[57,54,55,53,56]]