Add contribution instructions regarding LLM usage policy [skip ci]

CONTRIBUTING.md

@@ -0,0 +1,46 @@
+# Contributing
+
+Contributions are warmly welcome. Before you start, please read the entire document carefully.
+
+## AI Policy
+
+Do not use LLMs to generate substantial pieces of code. You need to fully understand every line of code you change. No exceptions.
+
+Commits, issues and pull requests must be made by humans. Co-authoring by AI is not allowed.
+Looking to create an issue on the project, such as a feature request or bug report? Focus on writing what you know for sure in simple terms. We'll figure it out from there.
+
+Usage of Gen-AI for visuals and audio is prohibited.
+
+The only exception is LLM-translated strings generated by the scripts/translator tool.
+
+## Basics
+
+Before committing, make sure to format your Rust code by running:
+
+```
+cargo fmt
+```
+
+You should also check for any compiler warnings too:
+
+```
+cargo check
+```
+
+or even better:
+
+```
+cargo clippy
+```
+
+## Large contributions
+
+If you want to submit a relatively big pull requests, it's highly advisable to create a separate issue request for it. That way we can discuss the approach and sort out the details first.
+
+## Adding translations
+
+Look at `scripts/translator/run.sh` for more information.
+
+## Licensing
+
+WayVR is developed and published under [GPL-3 License](https://www.gnu.org/licenses/gpl-3.0.en.html).

scripts/translator/run.sh

@@ -1,6 +1,32 @@
 #!/bin/bash
 
-# Do not run this script if you don't have at least 32 GiB of system memory.
+# This is the translation tool intended for machine-translating strings.
+#
+# ###########################
+# # Adding new translations #
+# ###########################
+# 
+# If you want to add a new language, follow these steps:
+#
+# 1. Include the language in `wlx-common/src/locale.rs`:
+#   - add your entry in `enum Language` 
+#   - add your entry in Language::code() function
+#   - add your language code in Language::add_codes() function
+# 
+# 2. Create a new template file
+#   - Add your language code in `scripts/translator/templates/<lang code>.json`
+#     (you can copy one of nearby files and adapt them accordingly)
+# 
+# 3. Modify this very file
+#   - Go to the bottom and add your language code entry here too
+#
+# 4. Run the script (it will hint you which LANG_PATHs are supported)
+# 
+# ############################################
+# # Running machine translations (local LLM) #
+# ############################################
+#
+# Do not run this script if you don't have at least ~16 GiB of free system memory.
 # Submitting PRs that lack translated strings is okay if you don't meet the
 # system requirements to run this script, or if you simply prefer not to; we are
 # regularly updating the missing translation strings anyway.
@@ -8,15 +34,18 @@
 # Base language: English (en.json)
 # 
 # This tool expects llama.cpp server being present (llama-server) at port 8080.
+# 
+# Currently we're using Gemma 4 for machine translations.
+# Model URL:
+# https://huggingface.co/unsloth/gemma-4-31B-it-GGUF
+# Recommended quantization: Q4_K_M
+# 
 
 set -e
 cd "$(dirname "$0")"
 
 bun install
 
-# model URL:
-# https://huggingface.co/unsloth/gemma-4-31B-it-GGUF
-# you can specify your own if you want to.
 if [ -z "${MODEL}" ]; then
 	export MODEL="gemma-4-31B-it-Q4_K_M"
 fi