Source Documents
Upload documents and let AI extract architectural knowledge
Source Documents let you upload existing documentation --- RFPs, architecture decision records, design docs, or any text --- and have Archflow's AI extract systems, connections, and documentation blocks automatically.
Overview
Instead of manually creating every system and connection, upload your existing documents and let the AI bootstrap your architecture project. The AI parses the content to identify:
- Systems --- Software components, services, and actors mentioned in the document
- Connections --- Relationships and data flows between identified systems
- Doc Blocks --- Relevant knowledge snippets that become project-level context
Accessing Source Documents
From the project Dashboard, open the Source Documents tab to manage your uploaded documents.
Upload and Parse Workflow
1. Upload
- Click Upload and select a document file
- Supported formats include text files and markdown
- Files show size and upload timestamp
2. Parse
- Click Parse on an uploaded document (or use Parse All for batch processing)
- The AI analyzes the document content and extracts architectural elements
- Parsing status is tracked: Pending → Parsing → Parsed (or Error)
3. Review
- View the extraction summary showing discovered systems, connections, and doc blocks
- Click a document to open the Detail Modal with full extraction results
- Check accuracy before syncing to your project
4. Sync
- Click Sync to apply the extracted data to your project
- New systems and connections are created in your architecture
- Doc blocks are stored as project-level knowledge for AI context
- A Sync Preview Modal shows exactly what will be added before you confirm
5. Reparse
- If parsing failed or you've updated the source document, use Reparse Failed to retry
- Documents can be re-parsed as your project evolves
Document Status
| Status | Description |
|---|---|
| Pending | Uploaded, waiting to be parsed |
| Parsing | AI is currently analyzing the document |
| Parsed | Successfully extracted architectural knowledge |
| Synced | Extracted data has been applied to the project |
| Skipped | Document was skipped (e.g., irrelevant content) |
| Error | Parsing failed --- can be retried |
Extraction Summary
After parsing, each document shows an extraction summary:
- Systems found --- Number of systems/services identified
- Connections found --- Number of relationships between systems
- Doc Blocks found --- Number of knowledge snippets extracted
Use Cases
Project Bootstrapping
Upload existing architecture documentation to quickly populate a new project with systems and connections, then refine from there.
Knowledge Capture
Upload meeting notes, RFPs, or design documents to capture architectural decisions and context as doc blocks.
Architecture Migration
When migrating from another tool, export documentation as text and import it as source documents to extract the architecture.