{% extends "base.html" %} {% block title %}ArtificeOCR{% endblock %} {% block content %}
  1. 1SourceChoose Tropy pages
  2. 2ProcessOCR and clean
  3. 3ReviewEdit beside the scan
  4. 4ReturnAdd notes or compile

Tropy source

Pages to transcribe

0 files
Start from your archiveChoose lists, items, and individual pages from a read-only Tropy project.
Or add files outside Tropy
Drop image files here or click to browse
File OCR Cleanup Title Translate Conf Time Status
Stages
Output

Progress

0%
Source scan

(no document selected)

No document selected

RAW OCR
(not run)
CLEANED
(not run)
TRANSLATED
(not run)

History

Runs

StartedStagesFilesFailedElapsed

Documents

FileStateLanguageConfReview
Source scan

(no document selected)

No document selected

RAW OCR
(not run)
CLEANED
(not run)
TRANSLATED
(not run)

OCR settings

Choose how Artifice reads and processes your documents. Changes apply to new runs after you save them.

Local models

Automatic detects Ollama and LM Studio, then chooses a suitable installed model when a run starts.

Choose a local service to see its installed models.

Automatic lets Artifice choose a suitable installed model. Connection addresses are discovered for you when possible.

Other providers and connection addresses Optional

Processing defaults

Safe starting values for mixed historical documents. Change them only when the collection needs it.

Advanced model limits Less common controls

Tropy and folders

Projects are read directly; reviewed text returns through Tropy’s Developer API. Artifice never edits the project database.

Approved folders

Artifice can read and write only inside locations you approve.

    Diagnostics

    Connection checks run automatically when Settings opens and can be repeated before a large collection.

    Checking local services…

    No changes
    Ready
    {% endblock %}