免費 / 開源在本機或雲端執行與打造開源 AI 模型
Agenta 每月有 45,760 次造訪,月增率 +29.5%,定價模式為免費增值。Agents-Flex 每月有 5,088 次造訪,成長率 +202.3%,定價模式為 Free / Open Source。

開源 Java 框架,用於打造 AI 智慧代理與 LLM 應用
| 面向/指標 | ![]() | ![]() |
|---|---|---|
| 官方網域 | agenta.ai | agentsflex.com |
| 定價模式 | 免費增值 | Free / Open Source |
| 每月造訪次數 | 4.58萬 (45,760) | 5088 (5,088) |
| 月流量變化量 | +1.04萬 (+29.5%) | +3405 (+202.3%) |
| 主要市場與市佔 | 🇺🇸 美國 (23.1%) | 🇨🇳 中國 (100%) |
| 每月直接造訪 | 1.72萬 | 426 |
| 使用者評分 | 4.4 | 4.6 |
| 支援平台 | 網站 | 網站 |

用於建構、評估與部署 AI Agent 的開源工作空間
從範本或空白開始建立 Agent,為其配置模型、技能與工具,並在聊天介面、Slack 或 Telegram 中執行。
透過 MCP 接入 Notion、GitHub、HubSpot、Slack、Linear、Drive 等,可在不重寫 Agent 的前提下更換工具或技能。
為 Agent 輸出評分、收集回饋,並檢視每次執行的耗時、Token 用量與成本紀錄,讓改進可被衡量。
自行決定哪些動作可無人值守執行、哪些需要審批,憑證則保存在 Agent 可用但不可讀的保險庫中。

開源 Java 框架,用於打造 AI 智慧代理與 LLM 應用
為任一支援的供應商設定 API Key 與模型名稱,再透過 chat() 呼叫,取得跨模型一致的介面。
AgentRunner 結合策略與中介層,在單一執行流中協調模型、工具與沙箱。
可重用、可版本化的能力定義,捆綁 SKILL.md、腳本與資源,可從本地目錄或遠端儲存探索載入。
依需求加入知識檢索、圖像生成與 OCR 流程,靈活混合同步與非同步執行方式。
免費 / 開源在本機或雲端執行與打造開源 AI 模型
免費 / 開源讓多個程式智能體在各自獨立的 worktree 中並行運作

免費 / 開源Mastra is an open-source TypeScript framework for building, deploying, and observing AI agents and workflows. It gives JavaScript developers typed agents, durable workflows, memory, MCP tool support, and built-in observability so prototypes can scale into production.
免費 / 開源Mastra is an open-source TypeScript framework for building, deploying, and observing AI agents and workflows. It gives JavaScript developers typed agents, durable workflows, memory, MCP tool support, and built-in observability so prototypes can scale into production.
Pinecone is a managed vector database and AI knowledge platform for building retrieval-augmented generation (RAG), semantic search, and agent memory. It stores vector embeddings at scale, makes writes searchable within seconds with no index tuning, and keeps p99 latency steady as data grows. Pinecone powers products for more than 10,000 customers and over a million developers worldwide.
Pinecone is a managed vector database and AI knowledge platform for building retrieval-augmented generation (RAG), semantic search, and agent memory. It stores vector embeddings at scale, makes writes searchable within seconds with no index tuning, and keeps p99 latency steady as data grows. Pinecone powers products for more than 10,000 customers and over a million developers worldwide.

免費 / 開源面向生產級 AI 的代理可觀測性、評估與持續改進平台

免費 / 開源Sim is the open-source AI workspace for building, deploying and governing AI agents. Create agents visually, in Chat or in code, ground them in a synced knowledge base, connect 270+ integrations, and control access, spend and deployment from one place — on Sim Cloud or self-hosted.
免費 / 開源Sim is the open-source AI workspace for building, deploying and governing AI agents. Create agents visually, in Chat or in code, ground them in a synced knowledge base, connect 270+ integrations, and control access, spend and deployment from one place — on Sim Cloud or self-hosted.