paperwise

Access Required

Sign In

Create an account or sign in to continue.

Sign In

Use your existing account.

Sign Up

Create a new account.

Enter credentials to continue.

Intelligence

Model Configuration

Reusable provider connections, plus per-task model overrides for metadata, retrieval, and OCR.

Overview

Model config summary

One-glance view of which connection and model serve each LLM-backed task.

Task Connection Configuration Status
Metadata Extraction - Not configured
Search and Ask Your Docs - Not configured
OCR - Not configured
Providers

Model connections

Create reusable provider connections once, then assign them to specific tasks below.

Per-task assignment

Task models

Assign an explicit connection and model for each LLM-backed task. Leave the model blank to inherit the connection default.

Metadata Extraction

Title, correspondent, dates, suggested tags.

Grounded Q&A

Powers Ask Your Docs answers and citations.

OCR

Reads scans, screenshots, and image PDFs.
OCR · Advanced

Image extraction settings

When OCR Engine is set to LLM, Paperwise sends rendered page images to the model. Local Tesseract is faster but less accurate on handwriting and low-contrast scans.

Auto-switch to OCR
Use OCR only when local text extraction returns less than 200 characters or looks like garbage.
Save to apply model routing.