Open Training Corpus · Free API
11,611 real-world GitHub issue-and-patch instances across 3,886 repositories and 7 languages (Python, Rust, Go, C++, PHP, Java, JavaScript). Query the corpus over HTTP, connect an MCP client, or download all of it in one 27.5 MB file.
No API key. No registration. No payment. Requests to /v1/* are rate limited to 120 per minute per IP.
PolyPatch-11k is a training and mining corpus: each instance pairs an issue description with the test changes that define the task and the test ids those changes touch. Test ids are derived statically — no test was executed (test_ids_verified is false on every instance), so this is not an execution-verified benchmark. The gold solution patch is deliberately withheld from every response; provenance.source_url resolves to the originating GitHub commit.
| Endpoint | Purpose |
|---|---|
| POST /v1/instances/fetch | Sample instances, filtered by language, repo, ci_annotated, min_tests |
| POST /v1/instances/search | BM25 full-text search over problem statements |
| GET /v1/instances/{id} | Retrieve one instance, e.g. /v1/instances/tokio-rs__tokio-6978 |
| GET /v1/schema | Field contract, withheld columns, caveats, license, citation |
| GET /v1/count | Corpus totals and per-language breakdown |
| GET /v1/bulk | 302 redirect to the full 27.5 MB zstd JSONL dump |
| POST /mcp | MCP server (Streamable HTTP, no auth): fetch_instances, search_instances, get_instance, get_corpus_stats |
curl -s -X POST https://polypatch.lateos.ai/v1/instances/search -H 'content-type: application/json' -d '{"q": "race condition", "language": "go", "limit": 5}'
Or point any MCP client at https://polypatch.lateos.ai/mcp — no auth headers needed.
The compilation and derived columns are MIT-licensed. Each instance embeds code from its upstream repository under that repository's own license — check provenance.source_url before redistributing or training on an individual instance. Derived from SwingBench-data (SwingArena, arXiv:2505.23932).