Add the documents you want to compare — proposals, contracts, papers, resumes. Each is parsed to plain text and kept as a whole document; the comparison graph fans out one parallel extraction per document (no chunking, no index).
Upload documents
Drag & drop files here, or browse
.pdf · .docx · .txt · .md
…or paste one document
No documents yet — add at least two above, then head to Compare.
What are you comparing?
Fields to extract
Each field is typed — the extraction node returns structured output against this schema, one record per document, all in parallel.
Documents to compare
Synthesis
Comparison table
Settings are locked
Enter the settings password to view or change the model & Bedrock connection.
Configuration
Stored server-side in SQLite (app_config.db) — never in the browser.
Model & connection
anthropic.* models take the native tool-use path for typed extraction; other families (e.g. qwen.*) use the OpenAI-compatible endpoint with a strict-JSON fallback.