
开源的桌面 AI 劳动力,替你完成浏览器、文件与应用任务
协调智能体会将目标拆解为步骤,并调度浏览器、终端、文档与研究等专业智能体,在本地工作区内并行作业。
桌面应用开源,可在本地运行或自托管于自有基础设施,使用自己的密钥,让数据与密钥始终受你掌控。
可接入云服务商、企业网关或本地模型并随时切换,无需重建工作流;支持自带 API Key 或免费本地模型。
跟踪每个智能体的搜索、来源与文件变更,比较版本,然后审批、修改或追问后再采纳输出。

面向文档抽取、结构化 JSON 与 OCR 的前沿视觉语言模型
NuExtract 与 NuExtract3 是专为结构化信息抽取和 OCR 构建的视觉语言模型,在同规模模型中基准成绩领先于通用模型。
把 PDF、扫描件和混合格式文档转成可直接被下游系统消费的带类型 JSON 对象或整洁 Markdown。
基于 LLM 标注数据预训练的专用编码器系列,适用于实体识别、分类与打标流水线。
可以先在托管平台试用,再在数据驻留或合规要求下把同一批模型部署到自己的基础设施。
免费 / 开源具备集成式 AI 代理节点的开源工作流自动化平台,但缺乏 Eigent 的原生桌面终端与文件操作功能。

免费 / 开源Open-source personal AI assistant running locally, but focuses more on messaging integrations than heavy terminal orchestration.
免费 / 开源Open-source personal AI assistant running locally, but focuses more on messaging integrations than heavy terminal orchestration.
Generates business apps and automated agents from text prompts, but functions primarily as a collaborative cloud workspace.
用于网页与数据工作流的可视化无代码 AI 代理构建器,但在云端而非本地桌面执行。

No-code platform for building specialist multi-agent workforces, but lacks local desktop execution capabilities.
No-code platform for building specialist multi-agent workforces, but lacks local desktop execution capabilities.

免费 / 开源可在桌面上运行AI代理的本地协作团队,但开发者生态系比Eigent没那么成熟。

Upstage builds enterprise-grade large language models and document AI engines, including Solar LLMs, Document Parse, Information Extract and an agent studio for document workflows.
Upstage builds enterprise-grade large language models and document AI engines, including Solar LLMs, Document Parse, Information Extract and an agent studio for document workflows.

免费 / 开源Veryfi is a document intelligence platform built around OCR and multi-modal extraction APIs that turn invoices, receipts, bank statements, checks and 100+ other document types into structured data, with on-device capture SDKs and real-time fraud detection included in the same API responses.
免费 / 开源Veryfi is a document intelligence platform built around OCR and multi-modal extraction APIs that turn invoices, receipts, bank statements, checks and 100+ other document types into structured data, with on-device capture SDKs and real-time fraud detection included in the same API responses.