Smart Import
Smart Import is the fastest way to populate your knowledge base. Instead of manually creating files one by one, you can upload documents, connect Google Drive, import from Notion, crawl a website, or pull in your email history. The AI pipeline reads your content, extracts the useful information, cleans up formatting, and organizes it into your knowledge base.
How It Works
The import process has three stages:
- You provide the source - upload files, select a Drive folder, connect Notion, enter a website URL, or choose an email account and date range
- AI processes the content - each file is read, cleaned up, and organized. Related documents are merged. Duplicates are detected.
- Results are added to your knowledge base - processed content goes directly into your knowledge base, ready for Jarvis to search
The whole process is automatic. You start the import, and the results appear in your knowledge base when processing is complete.
The Smart Import page showing the import wizard with file upload, Google Drive, and other source options
File Upload
The most straightforward way to import - upload files directly from your computer.
Supported formats: PDF, DOCX, XLSX, HTML, CSV, Markdown, plain text, ZIP archives, and code files (.py, .js, .ts, and more)
Go to Smart Import
Navigate to Knowledge Base and open the Smart Import tab.
Upload your files
Drag and drop files onto the upload area, or click to browse. You can upload up to 200 files at once (100MB total). ZIP files are automatically extracted.
Select a target scope
Choose which scope the imported content should belong to - usually "general" unless the content is client-specific.
Start the import
Click Start Import. You will be taken to the Jobs tab where you can watch the progress.
ZIP files are a great way to import a whole folder of documents at once. The import automatically filters out system files (like macOS resource forks) and skips files that are not useful as knowledge content.
What Happens to Each File Type
The AI pipeline treats different file types differently:
- Markdown and text files pass through with minimal processing - the content is already clean
- PDFs and Word documents get cleaned up - formatting artifacts are removed, structure is preserved
- Spreadsheets are converted to readable text format
- Code files are stored as searchable reference material - no AI rewriting
- Email threads are distilled into structured summaries with key details, people involved, and resolutions
Google Drive
Connect your Google Drive to import entire folders of content.
Click Connect Google Drive
On the Smart Import page, select the Google Drive source and click Connect.
Authorize access
You will be redirected to Google to grant read-only access to your Drive files. DeskClone AI only reads files - it never modifies or deletes anything in your Drive.
Select a folder
Choose the folder you want to import. The import will include all files in that folder and its subfolders.
Start the import
Click Start Import. Google Docs and Sheets are automatically exported to a format the AI pipeline can process.
Google Drive access is separate from Gmail access. Connecting Drive for imports does not give DeskClone AI access to your email, and vice versa.
Notion
Import pages directly from your Notion workspace.
Connect Notion
On the Smart Import page, select the Notion source and provide your integration token.
Select pages
Choose specific pages or import your entire workspace. Notion pages are converted to Markdown format for processing.
Start the import
Click Start Import. The page tree is fetched and each page is processed through the AI pipeline.
Website
Import content directly from a website URL. The crawler reads the page, follows navigation and footer links to discover related pages, and imports them all as knowledge base articles.
Select Website
On the Smart Import page, select the Website source.
Enter the URL
Provide the URL of the page you want to import. The crawler will also discover and import linked pages from the site's navigation.
Start the import
Click Start Import. Each discovered page is processed through the AI pipeline and added to your knowledge base.
Website import is great for pulling in your existing help center, FAQ page, or product documentation. The crawler automatically finds related pages, so you only need to provide the starting URL.
Email History
Turn your past support conversations into knowledge base content. This is different from email monitoring (which creates issues from new incoming email). Email history import reads old threads and extracts the useful knowledge from them.
Select Email History
On the Smart Import page, choose the email history source and select the Gmail account to import from (you must have a Gmail account connected under Integrations first).
Choose a date range
Select the start and end dates for the threads you want to import. This controls which email conversations are included.
Start the import
Click Start Import. The AI reads your email threads, strips out reply chains and noise, and distills each thread into a structured knowledge article with a summary, key details, people involved, and the resolution.
Email imports count against your plan's Smart Import file budget (each thread counts as one file). Check your remaining budget on the import page before starting a large email import.
Cost Estimates
Smart Import uses AI to process your content, and the cost depends on the number and type of files. Before starting any import, you will see a cost estimate based on your selection.
The cost varies by file type:
- Text and Markdown files - minimal cost (content passes through with little processing)
- PDFs and documents - moderate cost (cleanup and formatting)
- Email threads - moderate cost (distillation and summarization)
A cost calculator is available on the import page to help you estimate before committing.
Tracking Import Progress
After starting an import, you are taken to the Jobs tab. Here you can see:
- Job status - pending, extracting (processing individual files), synthesizing (organizing and merging), or accepted (complete)
- Per-file progress - which files have been processed and their status
- Cost - the running cost of the import
Large imports (100+ files) may take several minutes to process. You can navigate away from the page and come back - the import continues in the background.
Reviewing Results
Once the import completes, the processed content is automatically added to your knowledge base. You can review the results in the Files tab to see what was created.
If the import merged several related source files into a single knowledge base article, you will see the combined result. The Sources tab shows which original files contributed to each knowledge base article, so you always have traceability.
What's Next
With your knowledge base populated, it is time to put Jarvis on your website. Head to Embed Widget to learn how to generate and install the chat widget.
Was this page helpful?