
為 AI 代理打造的網頁搜尋 API 與爬蟲
在大型獨立網頁索引上執行自然語言或關鍵字查詢,支援依網域、日期與類別篩選。
回傳清洗後的正文與 AI 擷取的亮點,讓代理取得最相關片段而非整頁內容。
可選快速、均衡或深度搜尋模式,讓助理既能即時回答,也能進行更有憑據的長時研究。
擷取公司資訊、人物等結構化結果,用於名單建構與資料管道。

專為文件擷取、結構化 JSON 與 OCR 打造的前沿視覺語言模型
NuExtract 與 NuExtract3 是專為結構化資訊擷取與 OCR 打造的視覺語言模型,在同等規模模型中基準成績領先通用模型。
把 PDF、掃描件與混合格式文件轉成可被下游系統直接消費的帶型別 JSON 物件或乾淨 Markdown。
以 LLM 標註資料預訓練的專用編碼器系列,適用於實體辨識、分類與標記流程。
可先在託管平台試用,再在資料駐留或合規要求下把同一批模型部署到自己的基礎架構。

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.

免費 / 開源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.
免費 / 開源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.

專項更強Crustdata indexes the public web into a live graph of 1B people and hundreds of millions of companies, exposed through a flexible REST API and an MCP server so agents can search, enrich and watch accounts.