Compare
Two tools by the company they keep. Every comparable product plots download curves against each other, which answers which is bigger — a total, and the question this project declines. This compares what each is declared alongside.
requests in 350 repositories · httpx in 292 · out of 2,343 sampled · computed 2026-08-24
Not measured as alternatives. They share context, but they also appear together often enough that nothing here says people choose between them. Absence of a measurement, not a finding that they are unrelated.
Declared with both
shared context| Package | with requests | with httpx |
|---|---|---|
| uvicorn | 80 | 113 |
| fastapi | 80 | 114 |
| pydantic | 105 | 142 |
| python-dotenv | 96 | 88 |
| aiohttp | 56 | 29 |
| beautifulsoup4 | 40 | 28 |
| pyyaml | 94 | 67 |
| python-multipart | 36 | 43 |
| playwright | 18 | 26 |
| aiofiles | 18 | 30 |
| pypdf | 13 | 17 |
| openai | 76 | 54 |
In requests’s neighbourhood only
Ranked companions of requests that httpx’s own list does not contain. That is a difference between two published neighbourhoods, not a claim that these never appear with httpx — both lists are truncated, and many of these do.
In httpx’s neighbourhood only
The same, the other way round: companions of httpx that requests’s list does not contain.
- pytest-asyncio 9 of 2,343
- opentelemetry-api 13 of 2,343
- apscheduler 13 of 2,343
- anyio 15 of 2,343
- posthog 8 of 2,343
- markdownify 8 of 2,343
- idna 9 of 2,343
- alembic 18 of 2,343
- sse-starlette 9 of 2,343
- psycopg2-binary 10 of 2,343
- pathspec 10 of 2,343
- curl-cffi 10 of 2,343