sto-warp 1.0.21

Star Trek Online screenshot recognition + ML training (standalone WARP/WARP CORE).

STO-WARP

Star Trek Online screenshot recognition + ML training, distributed as a standalone Python package.

What is this?

STO-WARP combines two tools in one package:

WARP (Weaponry & Armament Recognition Platform) — reads your in-game screenshots from the PC version of Star Trek Online and automatically fills in your SETS build. Detects equipment, traits, bridge officers, and ship information using computer vision and machine learning. Console screenshots (PlayStation, Xbox) are not supported.

WARP CORE — trainer interface built into WARP. Review and correct recognition results, confirm annotations, and contribute corrections to the community model so recognition improves for everyone.

Technology

• WARP — recognition pipeline (OCR + layout detection + icon matching). Reads STO screenshots and emits structured slot/item results.
• WARP CORE — Qt trainer UI for reviewing recognition output and fine-tuning the EfficientNet / MobileNetV3 models that drive WARP.

Install

The recommended way to install sto-warp is via our universal installation scripts, which will automatically configure an isolated environment using pipx.

Linux / macOS:

curl -sSL https://raw.githubusercontent.com/raman78/sto-warp/main/install.sh | bash

Windows (PowerShell):

Invoke-WebRequest -Uri "https://raw.githubusercontent.com/raman78/sto-warp/main/install.ps1" -OutFile "install.ps1"; .\install.ps1; Remove-Item "install.ps1"

On first launch sto-warp registers itself in your system menu — Start Menu on Windows, Launchpad / Spotlight on macOS, Activities / KRunner on Linux — so subsequent runs can be started by clicking the icon instead of opening a terminal.

For manual installation methods and more details, see INSTALLATION.md.

Data and models

On first run sto-warp downloads:

• ML models (icon_classifier.pt, screen_classifier.pt) from the Hugging Face hub.
• Cargo / ship / trait JSON from the community STOCD/SETS-Data repository.
• Community knowledge (knowledge.json), admin icon equivalence map, and a seed tarball of confirmed crops.

A small setup splash blocks the launcher on the very first start until all of the above finish, so the application opens with a complete data set. On every later run the cache is refreshed silently in the background. Everything is cached under ~/.config/warp/ (or $XDG_CONFIG_HOME/warp/ when set). Nothing is committed to this repository. See docs/WARP_GUIDE.md for the user-facing walkthrough and docs/SYNC_ARCHITECTURE.md for the technical view.

License

GPL-3.0 — see LICENSE.

Project docs

• INSTALLATION.md — install methods (pipx, native packages).
• CHANGELOG.md — release notes.
• PROJECT_CONTEXT.md — design context.
• CLAUDE.md — repository rules for AI-assisted development.
• docs/ — technical deep-dives (BOFF / trait detection, ML pipeline, …).