已比較的版本

索引鍵

  • 此行已新增。
  • 此行已移除。
  • 格式已變更。

...

English / 中文

DocChat: Ask DaVinci to answer questions based on the document content
錨點
English
English

Instantly enhance your document processing efficiency—DocChat opens a new chapter in intelligent interpretation!

With just one click, upload Word, PDF, PPT, Excel,請達哥基於這些 Doc 的內容回答問題或做資料的處理

面板
panelIconIdatlassian-light_bulb_on
panelIcon:light_bulb_on:
panelIconText:light_bulb_on:
bgColor#FFFAE6

上傳 Doc 後,請記得將要使用的 Doc 做勾選,可以勾選單份或多份檔案

\uD83D\uDCD8 Instructions

  • Step1:點擊左上角的 New chat 開啟新的對話

...

  • Step2:點擊右上角的 New File 上傳檔案,並勾選檔案。可以勾選單份或多份檔案

...

  • Step3:根據需求選擇適合的 Model 和 Conversation Style

...

展開
titleComparison of Davinci's model

Model

Token limit

Usage Scenarios

GPT-3.5-Turbo-4k

4000 Tokens

翻譯、寫作等
簡單的文書工作

GPT-3.5-Turbo-16k

16000 Tokens

GPT-4-Turbo-8k

8000 Tokens

資料分析、比對等
困難或複雜的工作

...

titleConversation Style

...

Style

...

Difference

...

Usage Scenarios

...

Precise

...

達哥回答較精準

...

資料分析、比對等

...

Creative

...

達哥回答較天馬行空

...

and TXT files. Leveraging its powerful analytical capabilities, DaVinci swiftly answers questions related to your document content, making data handling both easy and accurate.

Overwhelmed by the pressure of comparing technical documents, filtering resumes, or analyzing financial reports?

DocChat functions like your personal team of experts, gracefully handling a variety of documents, calmly and precisely performing comparisons and analyses, saving you valuable time and effort.

...

Capability/Limitation 

Suggested Model

GPT-4 series model

Supported File Types for Upload

Text and table content within *.pdf, *.ppt, *.docx, *.txt files.

Supported File Size

Text and table content: 50 MB

Supported Pages

pdf, docx, ppt: < 1000 pages

Supported File Types for OCR

Image text within *.pdf, *.ppt, *.docx files

Supported File Encoding

cp437, charmap, UTF-8

File Restrictions

  • Encrypted files are not supported

  • Text content in PDFs that cannot be selected after conversion is not supported

  • Document content that becomes garbled when copied from a PDF and pasted into a text editor is not supported

Support for Generating Suggestions

Yes

...

\uD83D\uDCD8 Instructions

  • Step1: Click on "+ Chat" to start a new conversation

    doc1.pngImage Added

  • Step2: Choose File and click on "+ File" to upload a file and select the file(s). You can check either a single file or multiple files.

    doc2.pngImage Added

    doc3.pngImage Added

  • Step3: Select the appropriate Model and Conversation Style based on your needs.

    doc4.pngImage Added

  • Step4: Enter your Prompt or useSmart Prompt Suggestion to automatically generate questions.

    doc5.pngImage Added

...

錨點
Chinese
Chinese
立即提升文件處理效率,DocChat 為您開啟智能解讀新篇章!一鍵上傳 Word、PDF、PPT、Excel,達哥憑借強大分析能力,快速回答文件內容相關問題,讓資料處理變得輕鬆而準確。
面對技術文件比對、履歷篩選、財報分析壓力山大?DocChat 功能如同您的個人專家團隊,優雅處理多樣文件,從容不迫地精準比較與分析,節省您寶貴的時間與精力。

...

使用建議/限制

建議使用模型

GPT-4 series model

支援上傳的檔案類型

*.pdf, *.ppt, *.docx, *.txt 中的文字與表格內容

上傳大小限制

文字與表格內容: 50 MB

上船文件頁數限制

pdf, docx, ppt: 少於 1000 頁

支援 OCR 的檔案類型

*.pdf, *.ppt, *.docx 中的圖片文字

檔案限制

  • 不支援已加密檔案

  • 不支援轉成 PDF 後無法選取的文字內容

  • 不支援 PDF 複製後貼至記事本為亂碼的文件內容

支援產生 Suggestion 

Yes

...

\uD83D\uDCD8 Instructions

  • Step1:點擊左上角的 + Chat 開啟新的對話

...

  • Step2:點擊右上角的 + File 上傳檔案,並勾選檔案。可以勾選單份或多份檔案

...

...

  • Step4:輸入你的 Prompt 或使用 Smart Prompt Suggestion (智能提問) 自動產生問題

...

Smart Prompt Suggestion (智能提問) : 上傳檔案後不知道如何針對檔案問問題時,可以勾選檔案並點擊 Suggestions,達哥智能提問功能會建議你十個關鍵問題,直接點擊問題就能直接提問。如果不滿意,還可以重新產生新的十個問題,直到有適合的提問為止

...

📚 Related Use Case

依標籤篩選 (依標籤顯示內容)
showLabelsfalse
max5
spacescom.atlassian.confluence.content.render.xhtml.model.resource.identifiers.SpaceResourceIdentifier@893
showSpacefalse
sortmodified
typepage
reversetrue
labelskb-how-to-article
cqllabel = "docchat-case" and type = "page" and space = "DW"

📋 Related Article

依標籤篩選 (依標籤顯示內容)
showLabelsfalse
max5
spacescom.atlassian.confluence.content.render.xhtml.model.resource.identifiers.SpaceResourceIdentifier@893
showSpacefalse
sortmodified
typepage
reversetrue
labelskb-how-to-article
cqllabel in ( "direct-chat-manual" , "webchat-manual" , "plugin-chat-manual" , "davinci-store-manual" ) and type = "page" and space = "DW" and parent = "524515"