Benchmarks — Hummingbird against FastAPI and Fastify
All series › Series

Swift on the Server with Hummingbird

A series on server-side Swift with the Hummingbird framework — from the first route to a deployable LLM gateway on Linux.

We are building an OpenAI-compatible LLM gateway in Swift with the Hummingbird framework. The gateway runs against a local MLX model on Apple Silicon and deploys as a fully statically linked binary on Linux — with streaming, API-key auth, rate limiting, Prometheus metrics, and distributed tracing. The demo project is swift-mlx-gateway on Codeberg, with a Git tag per article to follow along. Six articles plus the prologue.

  1. 0

    Swift on the Server — Hummingbird 2 and a Local LLM Gateway

    Prologue · Series: Swift on the Server with Hummingbird. Why Swift on the server is no longer a hobby project in 2026, how Hummingbird differs from Vapor and why we chose a local LLM gateway as our demo project.
    13. May 2026
  2. 1

    Hello Hummingbird — The Skeleton

    Article 1 · Swift on the Server with Hummingbird. We create the project, write Package.swift, set up the AsyncParsableCommand entry point, and build a minimal router with /healthz and /v1/models until curl gets an answer.
    14. May 2026
  3. 2

    Two Protocols, One Backend — Anthropic and OpenAI with Codable

    Article 2 of the Hummingbird series. We implement the Anthropic Messages API and the OpenAI Chat Completions API in the same gateway: Codable types for both protocols, two POST endpoints, input validation, mock inference.
    15. May 2026
  4. 3

    Real Inference — MLXClient and the Local Model

    Article 3 of the Hummingbird series. We connect the gateway to mlx_lm.server: MLXClient as a Swift actor, format conversion from Anthropic to OpenAI, error mapping with HTTP status codes. After this article, the gateway delivers real model responses.
    15. May 2026
  5. 4

    Streaming — Tokens Flow as They Are Born

    Article 4 of the Hummingbird series. We add streaming to both endpoints via Server-Sent Events: MLXClient.completeStream as AsyncThrowingStream, Anthropic events, OpenAI chunks, task cancellation. Tokens appear at the client as the model generates them.
    16. May 2026
  6. 5

    Auth and RequestContext — Closing the Gate

    Article 5 of the Hummingbird series. We add API key auth, token-bucket rate limiting, and spec-compliant error responses — using Hummingbird 2's generic RequestContext to carry typed middleware state without dictionaries.
    17. May 2026
  7. 6

    Observability and Linux Deployment — From swift run to systemd

    Article 6 of the Hummingbird series. We add Prometheus metrics via swift-metrics and swift-prometheus, a /metrics endpoint, MetricsMiddleware and TracingMiddleware. Plus: cross-compiling to a statically linked Linux binary via the Swift Static Linux SDK, a multi-stage Dockerfile, and a systemd unit for bare-metal Linux deployment.
    19. May 2026
  8. 7

    Benchmarks — Hummingbird against FastAPI and Fastify

    Article 7 of the Hummingbird series. We measure the LLM gateway against functionally equivalent implementations in FastAPI and Fastify — with a fair middleware stack, a deterministic backend stub, and oha as the load generator. The results are both what you expect and somewhat surprising.
    20. May 2026