diff --git a/axon/main.py b/axon/main.py
new file mode 100644
index 0000000..c92eb52
--- /dev/null
+++ b/axon/main.py
@@ -0,0 +1,65 @@
+from contextlib import asynccontextmanager
+
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+
+from axon.middleware.auth import APIKeyMiddleware
+from axon.middleware.rate_limit import RateLimitMiddleware
+from axon.middleware.request_context import RequestContextMiddleware
+from axon.models import MetricsResponse
+from axon.routers import crawl, extract, index, jobs, map, research, search, tavily, validate
+from axon.services.cache import RedisCache, get_cache
+from axon.services.metrics import metrics
+from axon.services.reranker import get_reranker, load_reranker
+from axon.services.index import init_index
+
+
+@asynccontextmanager
+async def lifespan(app: FastAPI):
+    await init_index()
+    load_reranker()
+    yield
+
+
+app = FastAPI(title="Axon Search API", version="1.0.0", lifespan=lifespan)
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+app.add_middleware(RequestContextMiddleware)
+app.add_middleware(RateLimitMiddleware)
+app.add_middleware(APIKeyMiddleware)
+
+app.include_router(search.router)
+app.include_router(extract.router)
+app.include_router(crawl.router)
+app.include_router(map.router)
+app.include_router(index.router)
+app.include_router(jobs.router)
+app.include_router(research.router)
+app.include_router(tavily.router)
+app.include_router(validate.router)
+
+
+@app.get("/health")
+async def health() -> dict[str, str]:
+    cache = await get_cache()
+    redis_status = (
+        "connected"
+        if isinstance(cache, RedisCache) and await cache.ping()
+        else "in-memory fallback"
+    )
+    reranker_status = "loaded" if get_reranker() is not None else "not loaded"
+    return {
+        "status": "ok",
+        "redis": redis_status,
+        "reranker": reranker_status,
+    }
+
+
+@app.get("/metrics", response_model=MetricsResponse)
+async def metrics_endpoint() -> MetricsResponse:
+    return MetricsResponse(**await metrics.snapshot())