---

name: interactive-apps-and-api description: "Guides LaTeX-OCR GUI, FastAPI, Streamlit, and Docker service workflows for interactive or deployed pix2tex OCR." disable-model-invocation: true metadata: disco-role: operating license: MIT

Interactive Apps and API

Use this sub-skill when the user wants the latexocr desktop GUI, screenshot capture, formatted output, the FastAPI service, Streamlit demo, Docker API container, or optional dependency troubleshooting for these interactive routes.

Quick Route

• For API, Streamlit, Docker, request formats, and ports, read references/api-and-ui.md.
• For desktop GUI behavior, screenshot tools, formatting, and desktop entries, read references/gui-workflow.md.
• For failures, read references/troubleshooting.md.
• Run scripts/check_api_dependencies.py before starting a server; it imports the FastAPI app without loading the model startup event or launching uvicorn.

API Summary

Install API dependencies:

pip install "pix2tex[api]"

Run the API app directly when model downloads and long-lived processes are allowed:

uvicorn pix2tex.api.app:app --host 0.0.0.0 --port 8502

The app exposes / for health, /predict/ for image uploads, and /bytes/ for raw byte uploads. Both prediction endpoints expect multipart field name file.

GUI Summary

Install GUI dependencies:

pip install "pix2tex[gui]"
latexocr

Use SCREENSHOT_TOOL=grim, SCREENSHOT_TOOL=spectacle, SCREENSHOT_TOOL=gnome-screenshot, or SCREENSHOT_TOOL=pil to override the screenshot backend when auto-detection picks the wrong one.

Boundaries

• Core model inference and image-quality guidance belongs in ../ocr-inference/SKILL.md.
• Dataset rendering, scraping, and training belong in sibling sub-skills.
• Do not launch GUI windows, desktop installers, uvicorn, Streamlit, or Docker unless the user explicitly wants a long-running or desktop-side-effect action.