Skip to main content
Pre-release

This release is a pre-release and may not be stable for production use.

RustAPI

RustAPI Logo

FastAPI-style Python Web Framework backed by a Rust (Tokio / Hyper) Engine — with embedded Rust Database Streaming, JWT & Argon2 Primitives, Tier 3 Rust-Native Routes, and a built-in MCP Server.

PyPI GitHub Repository License: MIT Python: 3.8+ Rust: 1.75+


📦 Key Open-Source Features Included

Feature Area Implementation Details
FastAPI Compatibility FastAPI & Request aliases, status codes, declared status_code (e.g. 201, 204), async/sync response_model with response_model_exclude/include, structured 422 validation error handling, FastAPI(lifespan=...) ASGI context managers, Depends(..., scope="function"), EventSourceResponse (SSE), jsonable_encoder, Query/Body/Path/Form/File, OAuth2PasswordBearer, HTTPBearer, APIKeyHeader, OpenIdConnect, get_openapi, app.frontend().
Maximum Rust Architecture Direct SQL routes (@app.get(path, sql=...)), lock-free TTL cache (cache_ttl=60), token-bucket rate limiter (rate_limit="100/minute"), SIMD vector similarity (cosine_similarity/vector_search), high-concurrency WebSockets pub/sub (ws.subscribe/app.broadcast), zero-copy disk streaming (await file.save_to()), and native SSE broadcast (sse="topic").
HTTP & Routing FastAPI-style @app.get, @app.post, APIRouter, Depends, app.dependency_overrides, response_model, reload=True.
Rust-Native Tier 3 Routes app.add_native_route() zero-GIL C-speed fast-paths (50,000+ req/sec).
Database Engine sqlx SQLite & PostgreSQL pool (app.connect_db()) with zero-copy JSON streaming (db.query_json()).
Power Primitives Native Rust encode_jwt() / decode_jwt(), Argon2 hash_password() / verify_password(), MiniJinja render_template(), SIMD vector search.
Response Classes HTMLResponse, JSONResponse, PlainTextResponse, RedirectResponse, StreamingResponse, EventSourceResponse.
Request Ergonomics req.json(), req.form, req.body, UploadFile (with save_to zero-copy disk streaming), WebSocket (with pub/sub channels).
AI / MCP Integration Embedded Model Context Protocol server (@app.tool(), @app.resource(), @app.prompt()) at POST /mcp, plus optional hardware-accelerated ONNX serving (pyrustapi[ai]).
Telemetry & Observability Terminal access logging (INFO: 127.0.0.1 - "GET /render HTTP/1.1" 200 - 0.45ms), Swagger UI (/docs), OpenAPI (/openapi.json).

⚡ Framework Standing & Multi-Framework Benchmarks

rustapi delivers 100% FastAPI developer experience (DX) while running on a high-performance Rust (Tokio/Hyper) core engine.

┌───────────────────────────────────────────────────────────┐
│                    Python Layer (DX)                      │
│ FastAPI Syntax • Pydantic Models • Routers • SSE • Docs   │
└─────────────────────────────┬─────────────────────────────┘
                              │  PyO3 C-Bindings
┌─────────────────────────────▼─────────────────────────────┐
│                    Rust Tokio/Hyper Core                  │
│ Async Multi-Thread IO • Tokio Workers • sqlx DB • SSE      │
└───────────────────────────────────────────────────────────┘

📊 Empirical Benchmark Results (oha Load Generator, $c=32$)

Measured under identical business logic and schemas across Python and Node.js frameworks:

Workload Scenario FastAPI (uvloop) Express.js RustAPI (v1.8.9) Fastify (Node.js) RustAPI Native Tier 3
Sync JSON (GET /json) 3,417 req/s 12,873 req/s 10,178 req/s (⚡ 3.0x vs FastAPI) 17,063 req/s 🚀 96,386 req/s (⚡ 7.5x vs Express)
Async JSON (GET /async_json) 5,430 req/s 13,638 req/s 8,117 req/s (⚡ 1.5x vs FastAPI) 17,122 req/s N/A
Validated POST (POST /items) 2,762 req/s 10,529 req/s 8,621 req/s (⚡ 3.1x vs FastAPI) 13,463 req/s N/A
Response Model (50 items) (GET /users) 1,916 req/s 10,072 req/s 1,928 req/s (⚡ Lower p95) 11,145 req/s N/A
High Concurrency (c=100) 6,088 req/s 12,637 req/s 7,571 req/s (⚡ 1.25x vs FastAPI) 15,523 req/s N/A

Detailed report & reproducible scripts: performance-reports/rustapi_vs_fastapi_vs_nodejs.md

🚀 Maximum Native Rust Architecture Benchmarks (PyRustAPI vs FastAPI Uvicorn, $c=50$)

Under high-concurrency production load with the maximum native Rust acceleration active:

Workload Scenario FastAPI (Uvicorn) PyRustAPI (Native Rust) Empirical Performance Boost Notes
DDoS / Rate Limit Shield (/rate-limit) 2,092.4 req/s (24.0 ms) 62,723.6 req/s (0.79 ms) 🚀 +2,897.6% (30x faster) Handled in Tokio before Python wakes up
SIMD Vector Search (10k vectors × 128 dim) 247.45 ms (Python loop) 23.36 ms (AVX2 registers) 🧠 +959.0% (10.5x faster) Native SIMD vector distance
Database Query (/sql SQLx vs SQLite) 258.2 req/s (201.5 ms) 1,524.0 req/s (32.8 ms) ⚡ +490.2% (6x throughput) Direct socket wire-protocol streaming
Standard JSON Response (/json) 2,609.5 req/s (19.2 ms) 7,732.2 req/s (6.4 ms) 🔥 +196.3% (3x faster) Hyper Ingress + PyO3 C-bindings
In-Memory Cache Hit (/cached) 2,394.2 req/s (20.9 ms) 5,754.9 req/s (8.6 ms) 💾 +140.4% (2.4x faster) Lock-free DashMap cache (sub-100µs)

Detailed report: performance-reports/maximum_rust_architecture_benchmark.md & guide: docs/md/maximum_rust_architecture.md


🏛️ The Maximum Native Rust Execution Model

  1. Tier 1 (FastAPI Drop-in): Standard Python @app.get(), async def, Pydantic models, and dependencies.
  2. Tier 2 (Hybrid Rust Acceleration): Standard Python handlers utilizing native Rust power primitives (db.query_json() SQLite/PostgreSQL streaming, rustapi.encode_jwt(), rustapi.hash_password(), MiniJinja templates, SIMD cosine_similarity).
  3. Tier 3 (Pure Rust Native Fast-Paths & Wire Streaming): Zero-GIL routes registered via @app.get(path, sql=...), @app.get(path, sse=...), or app.add_native_route("/health", ...) clocking up to 96,000+ req/sec with sub-millisecond latencies.
  4. Token-Bucket Shield & In-Memory TTL Cache: Sub-50µs 429 DDoS mitigation and sub-100µs RAM cache hits handled completely in Rust.
  5. Zero-Copy Disk Streaming: High-throughput file uploads (await file.save_to(path)) streamed directly to disk at flat 15MB memory.

🔄 Lifespan Context Manager (ML Models, DB Pools, Resource Cleanup)

RustAPI supports the modern FastAPI @asynccontextmanager lifespan protocol across both ASGI and native Tokio app.run():

from contextlib import asynccontextmanager
from rustapi import FastAPI

ml_models = {}

@asynccontextmanager
async def lifespan(app: FastAPI):
    # Startup: Load ML model or connect to database before handling requests
    print("Loading application resources...")
    ml_models["hero_model"] = "Loaded Model Instance"
    yield
    # Shutdown: Clean up resources and release memory on server exit
    print("Cleaning up resources...")
    ml_models.clear()

app = FastAPI(lifespan=lifespan)

@app.get("/")
async def read_root():
    return {"model": ml_models.get("hero_model")}

if __name__ == "__main__":
    app.run(host="127.0.0.1", port=8000)

⚡ 5 Proven Ways to Maximize Performance

  1. Pre-load Heavy Resources in lifespan: Load models and connection pools once during startup rather than in request handlers.
  2. Zero-Copy SQL with db.query_json(): Stream database query results directly to client sockets without allocating Python dictionary objects.
  3. Declare response_model for Compiled Serialization: RustAPI compiles Pydantic TypeAdapter instances once, boosting serialization by up to 64.5%.
  4. Offload Health Checks to Tier 3 Native Routes: Use app.add_native_route("/health", ...) for ultra-high-frequency endpoints (96,000+ req/s).
  5. Scale Multi-Process Workers: Set workers=os.cpu_count() in app.run() for multi-core parallelism.

📚 Documentation Reference

Detailed documentation is available in the docs/ directory and project reference:


🛠️ Local Development & Testing

Build from source with maturin:

# Clone the repository
git clone https://github.com/rajboopathiking/rustapi.git
cd rustapi

# Build release wheel with Maturin
pip install maturin
maturin develop --release

# Run full test suite
pytest tests/ -v

PYPI Package

pip install pyrustapi

📄 License

Distributed under the MIT License.

Metadata

Release files for pyrustapi 1.9.0b6

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for pyrustapi 1.9.0b6
File Size Uploaded Provenance
pyrustapi-1.9.0b6.tar.gz 3.6 MB Without provenance Details

Built distribution (wheel)

Table of built distributions (wheels) for pyrustapi 1.9.0b6
File Interpreter ABI Platform
pyrustapi-1.9.0b6-cp38-abi3-macosx_10_12_x86_64.whl CPython 3.8 abi3 macOS 10.12+ x86-64 Details

Total release size: 7.9 MB

Release files / pyrustapi-1.9.0b6.tar.gz

Download URL pyrustapi-1.9.0b6.tar.gz
Size 3.6 MB
Tags Source
SHA-256 checksum
How to use checksums
0ad142cad90617e5044f0533d157bd5acbbfc755038e7300c5750016c36c65d1
BLAKE2b-256 checksum
How to use checksums
7cafc390ae7797bc4abc611afa6dd120f4530162b270ac70bac69a0dc3eb6df0
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.11.9

Release files / pyrustapi-1.9.0b6-cp38-abi3-macosx_10_12_x86_64.whl

Download URL pyrustapi-1.9.0b6-cp38-abi3-macosx_10_12_x86_64.whl
Size 4.3 MB
Tags CPython 3.8 abi3 macOS 10.12+ x86-64
SHA-256 checksum
How to use checksums
45b42e096f1b8d7b654ef5b5e669014e983f2be3247ae3efd28431c296ab099e
BLAKE2b-256 checksum
How to use checksums
4f41cc3c1b33f5515030c8a275c02268e4102f68245ff8bdc36cac9ed383846e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.11.9

Security for pyrustapi 1.9.0b6

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page