Knowledge
The Knowledge section provides documentation about Rhesis’s core concepts and methodologies for AI testing and evaluation.
What are Sources?
Sources are content that you add to Rhesis to provide context and domain knowledge for AI testing. They serve as the foundation for generating relevant, domain-specific test cases.
Source Types
Rhesis currently supports:
- Document: Uploaded files (PDF, Word, Excel, PowerPoint, text files, etc.)
- Tool: Pages imported from MCP servers (e.g., Notion pages)
Future Support: Additional source types including Website, API, Database, Code, and Manual are planned for future releases.
Supported File Formats
When adding documents, Rhesis supports a wide range of file formats:
Office Documents:
.docx- Microsoft Word documents.pptx- Microsoft PowerPoint presentations.xlsx- Microsoft Excel spreadsheets
Documents:
.pdf- PDF files.txt- Plain text files.csv- Comma-separated values.json- JSON data files.xml- XML documents.html,.htm- HTML web pages
Archives:
.zip- Compressed archives (contents are extracted)
E-books:
.epub- Electronic book format
How Sources Work
Adding Sources
You can add sources in two ways:
- Upload Documents: Upload files directly from your computer
- Import from MCP: Import pages from connected MCP servers (see MCP documentation)
Storage and Processing
When you add a source:
- Content Storage: The content is securely stored in your organization’s storage
- Content Extraction: Text content is automatically extracted using Markitdown
- Metadata Creation: Source information (type, size, date added) is stored
- Source Record: A source record is created with the extracted content and metadata
You can then view details, browse extracted content, download the original file, edit metadata (title, description, tags), and add comments.
Importing from MCP
To import sources from MCP servers:
- Click “Import Source from MCP” on the Knowledge page
- Select an MCP tool (e.g., Notion connection)
- Enter a search query to find pages
- Select pages from the results
- Click “Import as Sources” to save them
Search Errors: If you encounter an error when searching, try again. The search agent is non-deterministic and may occasionally fail, but retrying may resolve the issue.
Using Sources for Test Generation
Current Capabilities
Sources are currently used in the Knowledge section to:
- Store Domain Knowledge: Add content that contains domain-specific information
- Extract Content: Automatically extract and view text content from sources
- Organize Information: Categorize and tag sources for easy management
Coming Soon: Sources will be integrated with test generation to automatically create domain-specific test cases.