ConlangCrafter: Constructing Languages with a Multi-Hop LLM Pipeline

Researchers at Tel Aviv University and the University of Ljubljana introduced ConlangCrafter, a fully automated multi-hop LLM pipeline that constructs artificial languages with coherent phonology, grammar, lexicon, and translation capabilities. The system, detailed in a paper on arXiv, generates diverse conlangs using large language models and includes a dataset of 64 generated languages. The project is open-source under an MIT license.

Project Page: conlangcrafter.github.io http://conlangcrafter.github.io Paper: arxiv.org/abs/2508.06094 https://arxiv.org/abs/2508.06094 Dataset: huggingface.co/datasets/malper/ConlangCrafter https://huggingface.co/datasets/malper/ConlangCrafter — 64 generated languages We introduce a fully automated system for constructing languages conlangs using large language models. Our multi-stage pipeline creates coherent, diverse artificial languages with their own phonology, grammar, lexicon, and translation capabilities. - Install dependencies: pip install -r requirements.txt or: uv sync if using uv - Set up API keys — copy .env.example to .env and add keys for whichever APIs you will use: Google Gemini : GOOGLE API KEY — Google AI Studio https://aistudio.google.com/app/apikey OpenAI : OPENAI API KEY — OpenAI API Keys https://platform.openai.com/api-keys DeepSeek via Together : TOGETHER API KEY — Together AI https://api.together.xyz/settings/api-keys - Generate a language sketch default model: gemini-2.5-pro : python src/run pipeline.py or: uv run src/run pipeline.py Run python src/run pipeline.py --help to see all options. Key flags: python src/run pipeline.py \ --model gemini-2.5-pro \ --custom-constraints "The language has only 3 vowels" \ --temperature 0.8 \ --qa-disabled QA self-refinement loops are on by default; use this to turn it off To resume a previous run e.g. starting from grammar after phonology completed : python src/run pipeline.py --language-id <id --steps grammar,lexicon Supported models are: - Google Gemini e.g., gemini-2.5-pro , gemini-1.5-flash - OpenAI models e.g., o4-mini , gpt-4o , gpt-5 - DeepSeek via Together AI e.g., deepseek-ai/DeepSeek-R1 You can load pregenerated language sketches from our dataset https://huggingface.co/datasets/malper/ConlangCrafter in this pipeline's format with this script: python src/load hf languages.py Translation is not run by default. To translate into a generated language, run the translation step separately. By default it translates the 10 sentences in configs/sentences default.txt : python src/run pipeline.py --language-id <id --steps translation To translate a single custom sentence instead: python src/run pipeline.py --language-id <id --steps translation --translation-sentence "Hello, world " Pass --translation-sketch-update to feed new vocabulary and grammar rules introduced during translation back into the sketch for each subsequent sentence, expanding the language as translation proceeds constructive translation . This implementation includes minor improvements to the system used for results from our paper: QA loop : Degenerate outputs e.g. JSON instead of text are detected and skipped inline, rather than post-hoc rejection sampling. QA amend prompt : Prompt wording is slightly adjusted for consistency with our system. @article{conlangcrafter2025, title={ConlangCrafter: Constructing Languages with a Multi-Hop LLM Pipeline}, author={Morris Alper and Moran Yanuka and Raja Giryes and Ga{\v{s}}per Begu{\v{s}}}, year={2025}, eprint={2508.06094}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2508.06094} } This project is licensed under the MIT License — see the LICENSE /morrisalp/ConlangCrafter/blob/main/LICENSE file for details.