mure 5.2.0

Perform multiple HTTP requests concurrently – without worrying about async/await.

mure

This is a thin layer on top of httpx2 that allows you to perform multiple HTTP requests concurrently without having to worry about async/await.

mure means multiple requests, but is also the German term for a form of mass wasting involving fast-moving flow of debris and dirt that has become liquified by the addition of water.

(The photo was taken by Leo Wehrli and is licensed under CC BY-SA 4.0)

Installation

Install the latest stable version from PyPI:

pip install mure

Usage

Pass a list of dictionaries with at least a value for url and get a generator with the corresponding responses. The first request is fired as soon as you access the first response:

>>> import mure
>>> from mure.models import Resource
>>> resources: list[Resource] = [
...     {"url": "https://httpbin.org/get"},
...     {"url": "https://httpbin.org/get", "params": {"foo": "bar"}},
...     {"url": "invalid"},
... ]
>>> responses = mure.get(resources, batch_size=2)  # nothing fired yet
>>> for resource, response in zip(resources, responses):
...     print(resource, "status code:", response.status)
...
{'url': 'https://httpbin.org/get'} status code: 200
{'url': 'https://httpbin.org/get', 'params': {'foo': 'bar'}} status code: 200
{'url': 'invalid'} status code: 0

The number of requests fired at the same time will never exceed batch_size – this is protected by a semaphore.

HTTP Methods

There are convenience functions for GET, POST, HEAD, PUT, PATCH and DELETE requests, for example:

>>> resources = [
...     {"url": "https://httpbin.org/post"},
...     {"url": "https://httpbin.org/post", "json": {"foo": "bar"}},
...     {"url": "invalid"},
... ]
>>> responses = mure.post(resources)

Verbosity

Control verbosity with the MURE_LOG_ERRORS environment variable:

>>> import os
>>> import mure
>>> next(mure.get([{"url": "invalid"}]))
<Response(0, UnsupportedProtocol("Request URL is missing an 'http://' or 'https://' protocol."))>
>>> os.environ["MURE_LOG_ERRORS"] = "true"
>>> next(mure.get([{"url": "invalid"}]))
[2026-06-10 13:19:22,546] [ERROR] Request URL is missing an 'http://' or 'https://' protocol.
Traceback (most recent call last):
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_transports/default.py", line 101, in map_httpcore_exceptions
    yield
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_transports/default.py", line 392, in handle_async_request
    resp = await self._pool.handle_async_request(req)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpcore2/_async/connection_pool.py", line 199, in handle_async_request
    raise UnsupportedProtocol("Request URL is missing an 'http://' or 'https://' protocol.")
httpcore2.UnsupportedProtocol: Request URL is missing an 'http://' or 'https://' protocol.

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/home/severin/git/mure/mure/iterator.py", line 162, in _asend_request
    response = await session.send(_request, follow_redirects=session.follow_redirects)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_client.py", line 1582, in send
    response = await self._send_handling_auth(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ...<4 lines>...
    )
    ^
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_client.py", line 1610, in _send_handling_auth
    response = await self._send_handling_redirects(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ...<3 lines>...
    )
    ^
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_client.py", line 1645, in _send_handling_redirects
    response = await self._send_single_request(request)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_client.py", line 1679, in _send_single_request
    response = await transport.handle_async_request(request)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_transports/default.py", line 391, in handle_async_request
    with map_httpcore_exceptions():
         ~~~~~~~~~~~~~~~~~~~~~~~^^
  File "/home/severin/.local/share/mise/installs/python/3.14.4/lib/python3.14/contextlib.py", line 162, in __exit__
    self.gen.throw(value)
    ~~~~~~~~~~~~~~^^^^^^^
  File "/home/severin/git/mure/.venv/lib/python3.14/site-packages/httpx2/_transports/default.py", line 118, in map_httpcore_exceptions
    raise mapped_exc(message) from exc
httpx2.UnsupportedProtocol: Request URL is missing an 'http://' or 'https://' protocol.
<Response(0, UnsupportedProtocol("Request URL is missing an 'http://' or 'https://' protocol."))>

Caching

You can enable caching to avoid requesting the same resources over and over again:

>>> import mure
>>> from mure.cache import Cache
>>> resources = [
...     {"url": "https://httpbin.org/post"},
...     {"url": "https://httpbin.org/post", "json": {"foo": "bar"}},
...     {"url": "https://httpbin.org/post"},
... ]
>>> responses = mure.post(resources, cache=Cache.SQLITE)

This will make only two requests and use the hit from the cache for the last resource. The responses are stored in a local SQLite database .mure-cache.sqlite in the current working directory.

Note that you have to install the SQLite extras.

You can also use the in-memory storage with Cache.MEMORY. This cache only persists within the same function call, i.e. calling mure.post() twice will create two separate caches.

There is also a Cache.FILE which stores the responses on disk (in a folder .mure-cache in the current working directory).

[!NOTE] Caching does not respect any Cache-Control HTTP headers or something like that. It just writes all responses, including unsuccessful ones, into the cache and may reuse them instead of firing another request. There is also no TTL mechanism.