Skip to main content

m0serve

Realtime from a synchronous Python app, with no added infrastructure.

A WSGI/ASGI server written in Mojo. Point it at a Django, Flask, FastHTML or Starlette app and it serves it — no Mojo toolchain required, no compilation step, nothing to configure:

pip install m0serve
m0serve myproject.wsgi:application

What it adds over gunicorn or uvicorn is that a plain synchronous view can hold a Server-Sent Events stream or a WebSocket, by answering with two response headers — and reach every subscriber on every worker with one function call. No Channels, no Redis, no daphne, no Pushpin, no second process:

from m0serve import m0pub

def events(request):                    # an ordinary synchronous view
    r = HttpResponse(": connected\n\n", content_type="text/event-stream")
    r["M0-Hold"] = "stream"             # m0serve holds the connection open
    r["M0-Channel"] = "news"            # ...subscribed to this channel
    return r                            # Django's part in it ends here


def announce(request):                  # any view, command, or cron job
    m0pub.publish("news", "deploy finished")
    return HttpResponse("ok")

Run it with m0serve myproject.wsgi --realtime. The view runs first, with sessions and permissions in hand — which is where your auth belongs. Under gunicorn the two headers are ignored and the same view degrades to a short plain response, so adopting this is not a fork of your codebase.

The ten-minute quickstart goes from pip install to live multi-tab sync from one synchronous Django file: https://github.com/codetalcott/mojo-http/blob/main/QUICKSTART.md — every command in it is executed by CI on every pull request.

It also hosts several applications in one process, each in its native execution mode: m0serve --mount /=shop.wsgi --mount /app=live.asgi runs sync Django on handler threads and async FastHTML on an asyncio executor, behind one listener with one shutdown.

The protocol is detected from the object, so the same command serves WSGI and ASGI. m0serve --help lists the flags; --workers N, --threads N (on free-threaded CPython) and --blocking-threads N are the topology ones.

When it is up it prints one line per worker — 🔥 m0serve: myproject.wsgi: application on http://0.0.0.0:8000 (protocol=wsgi workers=4) — after the application has been imported, so "it printed" means "it is serving". For an orchestrator use --health-path /health or a TCP check rather than the log line.

If something will not start, m0serve --doctor myproject.wsgi prints a JSON report — the interpreter it resolved and the virtualenv it came from, the spec it discovered, the topology it settled on, and every startup check with the fix for the ones that failed — and exits with the code the server itself would have used. It binds nothing and imports nothing twice.

What is in the wheel

One compiled binary and the Mojo runtime it loads. No CPython — m0serve resolves libpython at run time from the interpreter it is installed beside, which is why there is no ABI tag and one wheel per platform covers CPython 3.10 through 3.14, free-threaded builds included. It has no Python dependencies and fetches nothing at install time.

Install it into the same virtual environment as your application, the way you would gunicorn or uvicorn; the console script points the embedded interpreter at that environment.

Platforms

platform status
macOS arm64 (Apple Silicon) supported
Linux x86_64 (glibc) supported
Linux aarch64 (Graviton, Ampere, arm64 Docker) supported
macOS x86_64 (Intel) not possible — Modular ships no Intel Mac toolchain
musl / Alpine, Windows not supported

The wheel's filename carries the exact macOS and glibc floors it was built against. On an older system pip declines to install it rather than installing something that crashes.

Status

Pre-1.0 and deliberately small: HTTP/1.1 only, no TLS, no HTTP/2 — terminate at a proxy, which is gunicorn's answer too. The API will change before 1.0.

The full capability matrix, with the CI gate behind each row, is docs/SPEC.md.

Source, documentation and issues: https://github.com/codetalcott/mojo-http

Licence

MIT, and the wheel redistributes third-party components under their own terms — the Mojo runtime (Apache-2.0 with LLVM Exceptions) and a fork of lightbug_http (MIT). Full text and attribution ship inside the wheel; see NOTICE.txt beside the installed package.

Metadata

Release files for m0serve 1.13.0

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

Built distributions (wheels)

Table of built distributions (wheels) for m0serve 1.13.0
File Interpreter ABI Platform
m0serve-1.13.0-py3-none-manylinux_2_35_x86_64.whl Python 3 none Linux glibc 2.35+ x86-64 Details
m0serve-1.13.0-py3-none-manylinux_2_35_aarch64.whl Python 3 none Linux glibc 2.35+ ARM64 Details
m0serve-1.13.0-py3-none-macosx_13_0_arm64.whl Python 3 none macOS 13.0+ ARM64 Details

Total release size: 4.5 MB

Release files / m0serve-1.13.0-py3-none-manylinux_2_35_x86_64.whl

Download URL m0serve-1.13.0-py3-none-manylinux_2_35_x86_64.whl
Size 1.7 MB
Tags Linux glibc 2.35+ x86-64 Python 3
SHA-256 checksum
How to use checksums
28a002b022e4be8f48f888c10ebfb13ef7c7e6513077ce58ffeba390d280df54
BLAKE2b-256 checksum
How to use checksums
4d9e74d700b0132569f182e26246a0784e1b630d579c2c82bfa02c5effcf60c1
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

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.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 9, 2026.

Transparency log

Release files / m0serve-1.13.0-py3-none-manylinux_2_35_aarch64.whl

Download URL m0serve-1.13.0-py3-none-manylinux_2_35_aarch64.whl
Size 1.5 MB
Tags Linux glibc 2.35+ ARM64 Python 3
SHA-256 checksum
How to use checksums
ae5556322191065aae257a6caed53d1d6ee611327ca069c8ec0e301b73468bf2
BLAKE2b-256 checksum
How to use checksums
6c58e8fb046d7b541add8ea09b056d0ab8200cad0aa3f5cba7b4311a75fd596f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

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.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 9, 2026.

Transparency log

Release files / m0serve-1.13.0-py3-none-macosx_13_0_arm64.whl

Download URL m0serve-1.13.0-py3-none-macosx_13_0_arm64.whl
Size 1.3 MB
Tags Python 3 macOS 13.0+ ARM64
SHA-256 checksum
How to use checksums
88bce6069b23eb53398bd4dd313e0b9f0002b077f8f73e991c74979fb3c39356
BLAKE2b-256 checksum
How to use checksums
236cd58ed675c705b4d69e4302bdb8267cd2a5075700b62b5fa464ff408e874e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

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.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 9, 2026.

Transparency log
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