diff --git a/Dockerfile b/Dockerfile
index e076ee8..0d854c7 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -26,7 +26,7 @@ COPY requirements.txt ./
 RUN python -m pip install --upgrade pip \
     && python -m pip install -r requirements.txt
 
-COPY gunicorn.conf.py main.py readme.md ./
+COPY gunicorn.conf.py main.py README.md ./
 COPY shazam_project ./shazam_project
 COPY web ./web
 COPY scripts ./scripts
diff --git a/readme.md b/README.md
similarity index 100%
rename from readme.md
rename to README.md
diff --git a/TODO.md b/TODO.md
index 2347b48..6ac1dea 100644
--- a/TODO.md
+++ b/TODO.md
@@ -128,7 +128,7 @@ Complete these main tasks in order. A main task may be ticked only after its sub
 - [ ] Record dataset size, clip length, hardware, operating system, network region, date, provider plan, and warm/cold conditions.
 - [ ] Add confidence intervals or an explicit limitation explaining why the benchmark is too small for them.
 - [x] Implement a fourth local constellation-map/hash-pair backend over a small library, with an index builder and benchmark integration. Real-world accuracy and speed comparison remain open until the corpus is recorded.
-- [ ] Replace `Test set accuracy | X / Y songs matched correctly` in `readme.md` with generated, reviewable results.
+- [ ] Replace `Test set accuracy | X / Y songs matched correctly` in `README.md` with generated, reviewable results.
 - [ ] Remove or substantiate the existing `~2.1 s` and `~3.4 s` performance values.
 
 ### Matcher correctness
diff --git a/evaluation/README.md b/evaluation/README.md
index f020630..f49ebbd 100644
--- a/evaluation/README.md
+++ b/evaluation/README.md
@@ -130,7 +130,7 @@ PowerShell:
 ```powershell
 .\.venv\Scripts\python.exe scripts/update_readme.py `
   --results evaluation/results/benchmark.json `
-  --readme readme.md
+  --readme README.md
 ```
 
 Bash:
@@ -138,7 +138,7 @@ Bash:
 ```bash
 .venv/bin/python scripts/update_readme.py \
   --results evaluation/results/benchmark.json \
-  --readme readme.md
+  --readme README.md
 ```
 
 Review the diff, JSON, and Markdown report before committing. The command must not be used for partial, synthetic, credential-incomplete, or provider-free runs.
diff --git a/scripts/update_readme.py b/scripts/update_readme.py
index 02b39dc..1a358d0 100644
--- a/scripts/update_readme.py
+++ b/scripts/update_readme.py
@@ -143,10 +143,10 @@ def update_readme(readme_path: Path, results: dict[str, Any]) -> None:
 
 def main() -> int:
     parser = argparse.ArgumentParser(
-        description="Import complete benchmark results into readme.md."
+        description="Import complete benchmark results into README.md."
     )
     parser.add_argument("--results", type=Path, required=True)
-    parser.add_argument("--readme", type=Path, default=Path("readme.md"))
+    parser.add_argument("--readme", type=Path, default=Path("README.md"))
     args = parser.parse_args()
     try:
         results = json.loads(args.results.read_text(encoding="utf-8"))
diff --git a/tests/test_reproducible_benchmark.py b/tests/test_reproducible_benchmark.py
index 18583b9..e6764ee 100644
--- a/tests/test_reproducible_benchmark.py
+++ b/tests/test_reproducible_benchmark.py
@@ -554,7 +554,7 @@ def forbidden_provider(*_args, **_kwargs):
     assert summary["missing_inputs"] == 1
     assert results["records"][0]["error_code"] == "missing_clip"
     assert results["records"][0]["status"] == "invalid_audio"
-    readme = tmp_path / "readme.md"
+    readme = tmp_path / "README.md"
     readme.write_text(
         "before\n<!-- BENCHMARK_RESULTS:START -->\nold\n<!-- BENCHMARK_RESULTS:END -->\n",
         encoding="utf-8",
@@ -800,7 +800,7 @@ def _replace_records(results: dict, records: list[dict], clip_count: int) -> dic
 
 
 def test_readme_update_refuses_incomplete_results_without_writing(tmp_path):
-    readme = tmp_path / "readme.md"
+    readme = tmp_path / "README.md"
     original = "before\n<!-- BENCHMARK_RESULTS:START -->\nold\n<!-- BENCHMARK_RESULTS:END -->\n"
     readme.write_text(original, encoding="utf-8")
     incomplete = {"metadata": {"complete": False}, "backend_summary": {}, "clip_count": 1}
@@ -832,7 +832,7 @@ def test_validator_rejects_inconsistent_denominator_even_when_metadata_claims_co
 
 
 def test_readme_update_imports_only_generated_complete_metrics(tmp_path):
-    readme = tmp_path / "readme.md"
+    readme = tmp_path / "README.md"
     readme.write_text(
         "before\n<!-- BENCHMARK_RESULTS:START -->\nold\n<!-- BENCHMARK_RESULTS:END -->\n",
         encoding="utf-8",