[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"jgKgHvrI7c":3},"\u003Cp align=\"center\">\n  \u003Cimg src=\"logo.png\" alt=\"Lean Pool logo\" width=\"240\">\n\u003C/p>\n\n# lean-pool\n\n[![Lean Action CI](https://github.com/LeanPool/lean-pool/actions/workflows/lean_action_ci.yml/badge.svg)](https://github.com/LeanPool/lean-pool/actions/workflows/lean_action_ci.yml)\n[![Documentation](https://img.shields.io/badge/docs-online-blue)](https://leanpool.github.io/lean-pool/)\n[![Exposition](https://img.shields.io/badge/exposition-online-8a4fff)](https://leanpool.github.io/lean-pool/exposition/)\n[![Source profile](https://img.shields.io/endpoint?url=https%3A%2F%2Fvilin97.github.io%2Flean-source-profiler%2Fpool%2Fbadge.json)](https://vilin97.github.io/lean-source-profiler/pool/)\n[![Zulip](https://img.shields.io/badge/Zulip-Lean_Pool-6492FE?logo=zulip&logoColor=white)](https://leanprover.zulipchat.com/#narrow/channel/619231-Lean-Pool)\n[![Semantic Search](https://img.shields.io/badge/semantic_search-Octo-2f80ed)](https://octo.axiomatic-ai.com/search?scopes=repo%3ALeanPool%2Flean-pool)\n[![License](https://img.shields.io/github/license/LeanPool/lean-pool)](LICENSE)\n[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.20513444.svg)](https://doi.org/10.5281/zenodo.20513444)\n[![arXiv](https://img.shields.io/badge/arXiv-2609.25199-b31b1b)](https://arxiv.org/abs/2609.25199)\n\nLean Pool sits between [`mathlib`](https://github.com/leanprover-community/mathlib4) and [`merely-true`](https://github.com/merely-true/merely-true), preserving Lean 4 formalizations that don't fit mathlib's scope. Instead of mathlib's high-bar human review, it relies on deterministic linters and LLM judgment, so it can grow faster while staying `sorry`-free and pinned to the latest Mathlib. Tau Ceti is also a pinned Lake dependency; import its modules directly instead of vendoring projects already in Tau Ceti or Mathlib. See [`MOTIVATION.md`](MOTIVATION.md) for the why, browse the API docs at \u003Chttps://leanpool.github.io/lean-pool/>, and explore each project's dependency graph and declarations in the [exposition site](https://leanpool.github.io/lean-pool/exposition/).\n\nSemantic search is also available via the [API](https://search.octo.axiomatic-ai.com/api/search).\n\nExplore compile costs in the [LeanPool source profile](https://vilin97.github.io/lean-source-profiler/pool/),\nfrom projects and files down to declarations and source lines. The\n[weekly profiling job](scripts/weekly-profile/README.md) runs on our Azure VM;\nthe badge shows the date of the latest completed recording.\n\n\u003C!-- BEGIN STATS -->\n**298** formalization projects · **10,977,403** lines of Lean\n\u003C!-- END STATS -->\n\n\u003Csub>(stats above are refreshed automatically by the [generated-metadata workflow](.github/workflows/notice.yml) — edit [`python/lean_pool/stats.py`](python/lean_pool/stats.py), not the numbers)\u003C/sub>\n\nSo far, projects have been added by hand: each is a suitable, permissively licensed (Apache-2.0 or MIT) Lean repository, bumped to the latest Lean and Mathlib, made to pass [CI](.github/workflows/lean_action_ci.yml) — it builds warning-free and clears Mathlib's linters, the style checker, and the repository quality gates (no `sorry`/`admit`, no axioms beyond `Classical.choice`/`propext`/`Quot.sound`, no `unsafe`/`partial`, file headers, size limits) — and an [LLM review](.github/REVIEW_RULES.md) of fit and significance, then merged.\n\nLLM reviews use GPT-6-Astra at `xhigh` reasoning effort through a privately operated Codex worker, without paid OpenAI API requests or an API fallback. Reviews also show an estimated dollar cost at official Standard API token rates, labeled separately from Codex quota billing. See [review operations](python/review-operations.md) for deployment and diagnostics.\n\nProject PRs also receive an advisory Greptile review, configured in [`.greptile/`](.greptile/), for cross-file integration, reusable abstractions, completeness, maintainability, and measured cost. It supplements rather than replaces the independent LLM verdict.\n\n### Getting started\n\nRequires Lean (via [`elan`](https://leanprover-community.github.io/install/), with the toolchain pinned in [`lean-toolchain`](lean-toolchain)) and Python 3.13+ with [`uv`](https://docs.astral.sh/uv/).\n\n```bash\nmake setup    # pull Mathlib oleans, build the whole pool (~1.5h), install Python tooling\n```\n\nTo work on a single project you don't need the whole pool built — see the\n[fast per-project build](CONTRIBUTING.md#dev-setup) in `CONTRIBUTING.md`.\n\nTo regenerate the preserved Zeta5 numerical certificates, see the\n[certificate reproduction guide](scripts/zeta5-certificates/README.md).\n\nThe [moving-sofa certificate recipe](python/lean_pool/sofa_certificates/README.md)\nregenerates its optimized certificate modules from pinned public inputs.\n\n### Contributing\n\nSee [`CONTRIBUTING.md`](CONTRIBUTING.md).\n\nThe PR author or a maintainer can comment `/profile` to request an advisory\ncompile-cost report. Failed or zero-phase timed runs show unavailable timing\nand are excluded from totals; errors and any successfully measured heartbeat\ncounts remain visible.\n\nAccepted PRs enter GitHub’s merge queue. CI checks the combined changes against\ncurrent main without repeatedly updating authors’ branches. Each project owns\nits YAML card and public `Imports.lean`; there is no shared import list to edit.\nConflicts in the same proof files still require a manual repair.\n\n### Credits\n\nCreated as part of the [UW Lean Hackathon](https://uw2026leanhackathon.github.io/) by [Vasily Ilin](https://github.com/Vilin97) and [Justin Asher](https://github.com/justincasher).\n\n### Difference from similar projects\n\n[Tau Ceti](https://github.com/TauCetiProject/TauCeti) is another approach to solve the same problem. The differences are:\n- Lean Pool accepts human-written projects, not just AI projects.\n- Lean Pool is not a unified library like mathlib. Most projects are independent of each other.\n- Lean Pool only accepts completed formalization projects.\n\n[Palomar Registry](https://palomar-registry.org/) is also similar to Lean Pool. The differences are:\n- Lean Pool maintains accepted projects.\n- Lean Pool provides tools like search and documentation.\n- Palomar is a registry, not a unified repository.\n\nProjects accepted to the Palomar Registry may be submitted to Lean Pool, and priority will be given to them.\n\n### Citation\n\nTo cite Lean Pool, use the [paper](https://arxiv.org/abs/2609.25199):\n\n```bibtex\n@misc{ilin2026leanpool,\n  title = {{Lean Pool}: An {AI}-Maintained Archive of Formalized Mathematics},\n  author = {Vasily Ilin},\n  year = {2026},\n  eprint = {2609.25199},\n  archivePrefix = {arXiv},\n  primaryClass = {cs.AI},\n  doi = {10.48550/arXiv.2609.25199},\n  url = {https://arxiv.org/abs/2609.25199}\n}\n```\n",1791742484222]