get_document_full_content
Return the complete raw text of a specific document from the project’s knowledge base. Use document names fromlist_knowledge_documents. Partial names are matched case-insensitively. MCP has access to all project documents (no agent scoping).
Use Cases
- Read the full content of an uploaded document
- Export or process document text in your application
- Get complete text when search snippets are insufficient
- Build document viewers or full-text interfaces
Parameters
The document name as returned by
list_knowledge_documents (e.g. “Product Guide.pdf”). Partial names are matched case-insensitively.Response (success)
Response (not found)
When the document is not found, the response includessuccess: false, an error message, and optionally available (array of up to 10 document names in the project).
Example Request
Example Response (success)
Example Response (not found)
Behavior
- Partial matching: Pass a substring of the document name (e.g. “Product” for “Product Guide.pdf”); matching is case-insensitive
- Full text: Returns all chunks in order, concatenated — unlike
search_knowledge, which returns only relevant excerpts - Scoping: MCP can access all project documents; no agent filtering
Error Handling
Best Practices
- Get names from list: Call
list_knowledge_documentsfirst to get exact or display names - Use partial names: You can pass “Product” instead of “Product Guide.pdf” if it uniquely matches
- Handle large text: Full document content can be large; stream or paginate in your UI if needed
Related Tools
- list_knowledge_documents: List document names to pass here
- search_knowledge: Semantic search over knowledge (returns excerpts, not full text)
