I've decided to to try out Pyrefly, Ty and Zuban for their language server features (type checking disabled where possible) last week and found that Zuban is the fastest (but unfortunately doesn't currently have the option to disable type checking). Ty comes next and Pyrefly was surprisingly slow to load (have to wait a few seconds before I can goto definition for example).
They all lack certain features vs basedpyright (what I was using) such as auto imports (Ty has experimental support), showing signature/doc when selecting autocomplete options (I think Ty does have this one) and some other features that I can't remember.
One feature that always existed in Jedi (and now also Zuban) is "goto declaration" in Python. It allowed me to goto the "import" instead of the original definition of a function/class which I'm surprised either isn't supported at all (pyright?) or would just do the same thing as "goto definition" (Ty), seems like an obvious oversight imho.
Edit: Also, I wish all these new tools give more importance to such IDE features as much as they do for type checking.
Hey, Pyrefly developer here, thanks for trying us out!
We're dedicated to providing a great IDE experience, though it does take some time. Please bother us on github / our discord if you have features you want - bug reports / community asks are our biggest priority.
- auto import is implemented in Pyrefly: it uses your pyrefly.toml project structure or falls back to your VSCode workspace (up to the first 2500 files). we're happy to fix it for your situation if you want to provide a reproduction
- speed: by far the biggest issue. your problem is likely related to [this](https://github.com/facebook/pyrefly/issues/360) but we need more information to speed it up. we're happy to work with you to to improve this if you're willing to provide a project structure
If ruff & uv have proved anything, it's that a tool that's effortless, a net-positive and fast will get adopted.
New typecheckers don't need to be perfect. They need to be good enough, easy to integrate and have low false positives. Sure, they will improve with time, but if feels like a pain then no one will pick it up. Python users are famously averse to tools that slow down their dev cycles, even if it means better long term stability
BasedPyright is popular because it comes built-in with Cursor and disappears into the background. I have a positive bias towards Astral figuring it out given their track record. But, none of these tools have reached the point of effortlessness just yet.
> New typecheckers don't need to be perfect. They need to be good enough, easy to integrate and have low false positives. Sure, they will improve with time, but if feels like a pain then no one will pick it up. Python users are famously averse to tools that slow down their dev cycles, even if it means better long term stability
I really don't agree. Sure, they don't need to be perfect but also keep in mind many codebases have already standardized on something like (based)pyright or mypy. So there's a migration cost. If your analysis has a lot of false positives or misses a lot of what those type checkers miss then there's little incentive to migrate. Sure, ty and pyrefly are much faster, but at the end of the day speed is only one consideration for a type checker.
No, I'm pretty sure they need to be perfect. If the tool's telling me bad information about types, first I'm going to lose a stupid amount of time to that wrong info, and then swear off the tool forever after I figure the tool was wrong.
I don't think that's an accurate metaphor. Seatbelts are expressly a runtime solution to car crashes, whereas Python's type checking is a) only done during development (not even during build), and b) completely reliant on third-party tools.
If you're looking for a car analogy, I would suggest comparing Python's typing system to installing speed cameras on the factory floor.
At least for Ty, not sure about the others, it is explicitly for type checking, exposed as an LSP. It’s not trying to compete with full LSP implementations. Most modern editors let you combine multiple LSPs, you shouldn’t think of them as using only one at a time
We actually do want ty to be a first-class LSP (i.e., a complete alternative to Pylance and others), and it already supports nearly all of the features you'd expect. I use it as my primary LSP today in lieu of Pylance!
Anyone know how this compares to 'ty': a new typechecker from Astral (uv/ruff team), also written in Rust and super fast. I had been waiting on it to reach beta, but would love to move to something faster than pyright sooner if possible.
Anyone struggling with slow mypy should really update to latest version. This years releases has focused on performance and it has payed off. Add the boost from latest Python versions to that and you can see 50% type checking improvements. Still far from a rust based tool, always something, all without changing your tool chain.
Hm, it doesn't seem to be dealing particularly well with imported packages that don't have type annotations. Seeing a bunch of "has no attribute" warnings. Some of the "substitute" annotations also seem to be wrong (e.g. asyncio in CPython has no annotations [in my installed version], but it's pulling some in from somewhere and they're not quite right…) It's also getting confused about lists and tuples in __slots__.
The advantage is type hints can be fixed without needing to release a new version of Python. The disadvantage is there's a lot of code in the standard library that doesn't really consider how to represent it with type hints, and it can be really tricky and not always possible.
I'm surprised to see so many people moving to pyrefly, ty, and zuban so quickly. I was going to wait until some time in 2026 to see which has matured to the point I find useful, I guess some users really find existing solutions actually unworkable.
Hmm. Presumably mypy and pyrefly use the same ones, but then I don't understand why pyrefly is complaining and mypy isn't:
ERROR Argument `Future[list[BaseException | Any]]` is not assignable to parameter `future` with type `Awaitable[tuple[Any]]` in function `asyncio.events.AbstractEventLoop.run_until_complete` [bad-argument-type]
--> xxx/xxx.py:513:33
|
513 | loop.run_until_complete(tasks.gather(*x, return_exceptions=True))
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(tbh this is rather insignificant compared to the noise from external packages without type annotations, or with incomplete ones… pyrefly's inferences at the existence of attributes and their types are extremely conservative…)
> Hmm. Presumably mypy and pyrefly use the same ones, but then I don't understand why pyrefly is complaining and mypy isn't:
> …where is it even puling "tuple[Any]" from…
Perhaps it's a bug in pyrefly, perhaps mypy or pyrefly is able to infer something about the types that the other isn't. I would strongly suggest checking their issues page, and if not seeing a report already report it yourself.
While there is an ongoing push to more consistently document the typing spec: https://typing.python.org/. It does not actually cover all the things you can infer from type hints, and different type hint checkers have decided to take different design choices compared to mypy and will produce different errors even in the most ideal situation.
This is one of the reasons why I am waiting for these libraries to mature a little more.
> it does not actually cover what rules you can check or infer from type hints
Indeed this is the cause of maybe 30% of the warnings I'm seeing… items being added to lists or dicts in some place (or something else making it infer a container type), and pyrefly then refusing other types getting added elsewhere. The most "egregious" one I saw was:
def something(x: list[str]) -> str:
foo = []
for i in x:
foo.append(i)
return "".join(foo)
Where it complains:
ERROR Argument `str` is not assignable to parameter `object` with type `LiteralString` in function `list.append` [bad-argument-type]
--> test.py:4:20
4 | foo.append(i)
Edit: now that I have posted it, this might actually be a bug in the .join type annotation… or something
Edit: nah, it's the loop (and the LiteralString variant of .join is just the first overload listed in the type hints)… https://github.com/facebook/pyrefly/issues/1107 - this is kinda important, I don't think I can use it before this is improved :/
I assume in your example if you update the foo declaration to the following it solves the complaint:
foo: list[str] = []
If so this a type checking design choice:
* What can I infer from an empty collection declaration?
* What do I allow further down the code to update inferences further up the code?
I don't know Pyrefly's philosophy here, but I assume it's guided by opinionated coding guidelines inside Meta, not what is perhaps the easiest for users to understand.
I agree, but as a type checker it is a subjective choice, whether to be explicit and not make assumptions or whether to infer from certain certain patterns as probably correct and not be noisy to the user. Very glad to see they plan to fix this.
I love the speed advantage of pyrefly over (based)pyright but so far it doesn't seem to highlight as much as pyright does, for example it doesn't catch unreachable code like this:
def fn(x: str):
if x is None:
x = "123" # pyright flags that as unreachable code, pyrefly does not
Autocomplete for modules also doesn't work for me yet:
import os
os. # I'll get `ABC, `Any`, `AnyStr`, `AnyStr_co`, `BinaryIO`, ...
Looking forward to have a fast language server for python though, pyright is way too slow for large projects.
Hey, Pyrefly developer here, thanks for trying us out! Thanks for bearing with us with these issues you're experiencing while we're still in alpha.
We're planning on adding unreachable code diagnostics soon (github issue [here](https://github.com/facebook/pyrefly/issues/1292)). These come for free with Pyright so we don't want to regress features.
I'm happy to help diagnose/fix your autocomplete issue: it should work on modules. If you want to provide details here, on [discord](https://discord.gg/Cf7mFQtW7W), or as a Github issue (github/discord preferred) we'll fix it for you + anyone else with the problem.
It's funny because in other typed languages a string can be null because of string being a reference type as opposed to a value type. I think Python type hints make more sense than C# does in this respect, which gives me a chuckle (two of my favorite languages and the untyped one makes more sense with types). That said, not sure if you already have but I would add a github issue / ticket reporting your issue to raise awareness to the devs.
If you use the nullable types compiler option in C# (which defaults to enabled for new projects) then you need to declare you string as string? for it to be nullable :)
A quibble: Python is strongly typed. It’s not un-typed in any reasonable sense of the word. It’s dynamically typed, though: type information is on objects, not on the names referring to them.
In some other typed languages, that is. E.g. in C++, OCaml, Haskell, F#, Kotlin, and Rust (a non-exhaustive list) you can have non-nullable strings, and other objects.
It’s worth noting that even though the runtime allows nulls (i.e. None) anywhere, Python type checkers do distinguish between optional and mandatory types.
It's not valid from a typing perspective, but python will let you. If you want to disregard types though then none of this matters anyway and you won't get much benefit from these tools
Ty has autocomplete for imports, but it's hidden behind a toggle right now. They are still working on it. They index all the modules and functions, so you can just type the function name and it will suggest the correct import and insert it.
Though, I'm happy with basedpyright and usually disable type checking. It's great to have so many options in language servers for Python now that pylance is locked to vscode.
Python is starting to feel a bit like JavaScript circa 2014. Remember grunt, gulp, webpack, coffeescript, babel? Now we've got pyright, mypy, pyrefly, black, ruff, ty, flake8, poetry, uv...
I used to find this kind of tooling explosion exhausting (and back then with JS it truly was), but generally speaking it's a good sign that the community is hungry to push things forward. Choice is good, as long as we keep the Unix spirit alive: small, composable tools that play nicely together and can be interchanged.
Interestingly besides typescript, javascript in 2025 is still super fragmeneted but by a bunch of well-polished tools that all do approximately the same thing. esbuild/vite(rollup)/trubopack(swc), prettier/biome/oxc, npm/bun/pnpm/yarn, bun/node/deno/worker-runtimes
It's just people refusing to use new tools. Prettier is so much slower than any Rust formatter, it's painful. `yarn install` takes forever. Why not switch to Bun? It's 5 minutes.
A lot of the tools you list here are composable, mypy checks type hints, black formats code, because of the purpose and design ethos of those two tools they would never think to merge.
So which is it that you want, to just reach for one tool or have tools that have specific design goals and then have to reach for many tools in a full workflow?
FWIW Astral's long term goal appears to be to just reach for one tool, hence why you can now do "uv format".
I think that's the point. Every now and then a language will have a small explosion of new tooling, and all you can really do is wait for it to blow over and see what tools people adopted afterwards, it feels like Python is going through a period like that at the moment.
How well does ty work these days? Last time I checked it still produced a lot of false positives and false negatives. (Which was very much expected, given its alpha/WIP status.)
Honestly i barely use it, just started to give it a try by adding types where it makes most sense and running with pre-commit. No issues so far. But, again, I'm a very light user.
I've tried pyrefly, ty, pyright, and basedpyright, with a large complex code base written using PyCharm, and _none_ of them do as good a job as PyCharm, particularly in discovering more complex type inheritances. It's a pity because in other respects Zed (which relies on these) is a worthy competitor to PyCharm (and much faster!) -- but the endless squiggly lines because pyrefly can't figure out the type, is annoying (and turning off the warnings is unhelpful).
Hopefully one of these will get up to PyCharm's level (my money would be on ty as Astral is kicking a* these days).
PyCharm has very basic type checking, though. It's not strict.
> pyrefly, ty, pyright, and basedpyright
All of them will complain 2-4x more about your code than PyCharm. I had more than 300 typing errors when I first opened my 20k LOC project in pyright that I wrote in Pycharm.
PyCharm works great when your code is not annotated. It infers types very well. But it won't complain in a lot of cases when your code is annotated.
I think the 2nd best IDE for Python for me is Visual Studio proper, not Code. I know it sounds crazy, but Microsoft actually put some effort into their Python integration. Again, its a 2nd best, but that still says a lot about everyone else's editors.
Nowadays I'm finding myself using Zed a lot more, so maybe the story will be that all these nice Rust based tools become baked in giving it super powers for Python.
I don't think these are designed to "discovering" complex type inheritances.
They are designed for code which are more or less fully typed, as opposed to PyCharm which cobbles together a bunch of heuristics to try to make sense out of untyped code. An admirable quest, but not one I'm personally interested in.
And their insistence on only supporting this approach drove my entire team away from using PyCharm.
(From shallowly observing notifications on the 20+ typehint related issues I'm subscribed to, they seem to have kinda turned around and working toward fully supporting the python type system finally - possibly by integrating with one of the third-party type-checkers)
Another one I recently discovered and that has a very active maintainer is "zuban" or zuban-ls. It has replaced jedi-language-server for me, and aims to replace mypy as well.
Speed is one of those "Quantity has a quality all its own" things. We use very fast tools in a qualitatively different way, even though all that changed was how long it takes in seconds.
It is interesting that nobody was writing these tools in C or in C++. There are obvious ergonomic reasons, but perhaps also it matters that Rust cares a lot more about types than either of those languages.
> It is interesting that nobody was writing these tools in C or in C++
This one's easier to explain. People interested in tooling for a specific language probably want to write that tooling in that language (hence pip, poetry, mypy, jedi, etc). Normally that would be the end if it, if Python wasn't 10-100x slower than a natively compiled language. And going from Python to Rust is an order of magnitude easier than going from Python to C or Python to C++, because the compiler is so good at identifying silly mistakes. Rust is just a friendlier language.
It's just very hard to write such systems in C/C++. Even all these Rust versions are segfaulting and panicking quite a lot. So many corner and edge cases that can be found in Python code, and the memory handling is also hard.
The author of Zuban started writing it back in 2020 or 2021, so it took him more than 4 years to complete it. And he is the author of Jedi, so he had prior experience already.
I get why they'd panic, but why have enough segfaults that you noticed? So, I went and read the code.
Zuban seems to have a bunch of scary "I'm not sure if this is correct" unsafe blocks, which to me would be a red flag. I mean, it's better that there's a comment expressing the doubt, but my experience is that if you're not sure whether it's correct, it's probably not correct.
While acknowledging the risk of causing a pile-on (which I don't want!), would it be possible to have a link to them or a description of what the unsafe blocks accomplish? I'm intrigued if they are for performance or API ergonomics, if they are due to limitations of the borrow checker, the stdlib or crate dependencies.
For anyone reaching for unsafe, there are in many cases either an existing API (split_at_mut comes to mind). For others, using zero-copy or bytemuck instead of unsafe is a good idea too.
None of that is to say "never write unsafe", unsafe existing is pretty much one of the reasons for Rust to be :)
Yes, but why they did not write it in a compiled language? Pyright is pretty slow in large code bases and takes a lot of RAM. Javascript can be faster than python in some cases, but Python is so easily extendable with C,C++, Cython, Rust. They could use Python with one of the compiled language.
I'm sorry, I can't take seriously any piece of software which decided to prefix the previous version's name with "based". I'm aware this is a me problem.
Hah. I love the name. It implies that whatever the original “pyright” was doing wasn’t keepin’ it real. This new version, it’s “based” so it must be somehow more “real” and “grounded” and “legit”.
All I know is it is much more strict about stuff than pylance was.
It's also horrible for fasle positives unless your project happens to be the exact same setup as the maintainers' - I had to turn off the actual type checking on it. I've since moved wholesale to the Ty alpha and it feels a hell of a lot smarter.
Sometimes dynamic Python idioms are incompatible with typed Python. I personally think that's fine, since I consider static typing a significant improvement overall.
This isn’t. They actually fixed that bug. Then they changed their minds and backed the fix back out again because they don’t think you should write Python that way:
> I think EAFP is a very unfortunate and ill-advised practice.
if "baz" in bar and "qux" in bar["baz"]:
foo = bar["baz"]["qux"]
...
else:
...
If this were a linter then I would accept that it is going to be opinionated. But this is not a linter, it’s a type checker. Their opinions about EAFP are irrelevant. That’s idiomatic Python.
yeah we all heard that story every 3 months with all the previous package managers. until there's adoption by an overwhelming majority of projects, it isn't really settled yet.
I’ve literally never heard that much buzz and excitement about Python tool before. And I’ve seen them all.
All of them had some big issue that prevented it from getting mainstream. Either it was slow, or didn’t work with existing workflow, or had complex configuration, or something that prevented gradual adoption.
uv is universally praised as the second coming Christ in Python world (and for a good reason). So no, I doubt there will be something else. Not only you need to be better than uv, you also need to have community momentum.
Sadly, the question with both this and ty is: Does it support pydantic? If not, then it's not really helpful for many people, and right now AFAIK neither supports pydantic.
Pydantic is probably the problem here, but it is what it is.
Python's type system is the problem here, which cannot support even the native dataclass pattern and needs custom plugins written for each type checker.
The type system’s alright. It just gets especially tricky when you’re trying to check code which won’t exist until you run it. For instance, suppose you wanted to load your own module from a database with something like
foo = eval(result)
It can’t know what you’re going to load until it actually does it.
Things which lean heavily into metaprogramming, typically ORMs or things like Pydantic, fall into that category. I can’t hold that against the type system.
I don't think this is really true, I think native data classes can do just about everything you need, it's just that you need some tooling around them to get the same as Pydantic offers. I'm pretty sure we are going to see better native dataclass support in Pydantic going forward.
I have used native dataclasses with pydantic starting with v1.5 and it’s only improved with the the newer 2.x releases.
Aside from basic inheritance and complex nested types, the pydantic ‘TypeAdapter’ is awesome for simply validating native dataclasses. It’s a little slow, but everything pydantic is =)
I've decided to to try out Pyrefly, Ty and Zuban for their language server features (type checking disabled where possible) last week and found that Zuban is the fastest (but unfortunately doesn't currently have the option to disable type checking). Ty comes next and Pyrefly was surprisingly slow to load (have to wait a few seconds before I can goto definition for example).
They all lack certain features vs basedpyright (what I was using) such as auto imports (Ty has experimental support), showing signature/doc when selecting autocomplete options (I think Ty does have this one) and some other features that I can't remember.
One feature that always existed in Jedi (and now also Zuban) is "goto declaration" in Python. It allowed me to goto the "import" instead of the original definition of a function/class which I'm surprised either isn't supported at all (pyright?) or would just do the same thing as "goto definition" (Ty), seems like an obvious oversight imho.
Edit: Also, I wish all these new tools give more importance to such IDE features as much as they do for type checking.
Hey, Pyrefly developer here, thanks for trying us out!
We're dedicated to providing a great IDE experience, though it does take some time. Please bother us on github / our discord if you have features you want - bug reports / community asks are our biggest priority.
- auto import is implemented in Pyrefly: it uses your pyrefly.toml project structure or falls back to your VSCode workspace (up to the first 2500 files). we're happy to fix it for your situation if you want to provide a reproduction
- signature/doc when selecting autocomplete options is a known bug [here](https://github.com/facebook/pyrefly/issues/1090
- go to declaration: I've created an issue for that [here](https://github.com/facebook/pyrefly/issues/1291), it should be quick.
- speed: by far the biggest issue. your problem is likely related to [this](https://github.com/facebook/pyrefly/issues/360) but we need more information to speed it up. we're happy to work with you to to improve this if you're willing to provide a project structure
If ruff & uv have proved anything, it's that a tool that's effortless, a net-positive and fast will get adopted.
New typecheckers don't need to be perfect. They need to be good enough, easy to integrate and have low false positives. Sure, they will improve with time, but if feels like a pain then no one will pick it up. Python users are famously averse to tools that slow down their dev cycles, even if it means better long term stability
BasedPyright is popular because it comes built-in with Cursor and disappears into the background. I have a positive bias towards Astral figuring it out given their track record. But, none of these tools have reached the point of effortlessness just yet.
> New typecheckers don't need to be perfect. They need to be good enough, easy to integrate and have low false positives. Sure, they will improve with time, but if feels like a pain then no one will pick it up. Python users are famously averse to tools that slow down their dev cycles, even if it means better long term stability
I really don't agree. Sure, they don't need to be perfect but also keep in mind many codebases have already standardized on something like (based)pyright or mypy. So there's a migration cost. If your analysis has a lot of false positives or misses a lot of what those type checkers miss then there's little incentive to migrate. Sure, ty and pyrefly are much faster, but at the end of the day speed is only one consideration for a type checker.
No, I'm pretty sure they need to be perfect. If the tool's telling me bad information about types, first I'm going to lose a stupid amount of time to that wrong info, and then swear off the tool forever after I figure the tool was wrong.
If typechecking is tongue on cheek, why even bother?
Because "all or nothing" is a bad idea.
If car crashes still kill people, why wear safety belts?
I don't think that's an accurate metaphor. Seatbelts are expressly a runtime solution to car crashes, whereas Python's type checking is a) only done during development (not even during build), and b) completely reliant on third-party tools.
If you're looking for a car analogy, I would suggest comparing Python's typing system to installing speed cameras on the factory floor.
At least for Ty, not sure about the others, it is explicitly for type checking, exposed as an LSP. It’s not trying to compete with full LSP implementations. Most modern editors let you combine multiple LSPs, you shouldn’t think of them as using only one at a time
We actually do want ty to be a first-class LSP (i.e., a complete alternative to Pylance and others), and it already supports nearly all of the features you'd expect. I use it as my primary LSP today in lieu of Pylance!
Do you assign different responsibilities to different LSP servers when there multiple I suppose?
Anyone know how this compares to 'ty': a new typechecker from Astral (uv/ruff team), also written in Rust and super fast. I had been waiting on it to reach beta, but would love to move to something faster than pyright sooner if possible.
Pyrefly vs. ty: Comparing Python’s Two New Rust-Based Type Checkers (2025-05-27) https://blog.edward-li.com/tech/comparing-pyrefly-vs-ty/
HN discussion of above: https://news.ycombinator.com/item?id=44107655
How Well Do New Python Type Checkers Conform? A Deep Dive into Ty, Pyrefly, and Zuban (2025-08-29) https://sinon.github.io/future-python-type-checkers/
amazing reply. Thanks!
Anyone struggling with slow mypy should really update to latest version. This years releases has focused on performance and it has payed off. Add the boost from latest Python versions to that and you can see 50% type checking improvements. Still far from a rust based tool, always something, all without changing your tool chain.
Anyone here know what the ideal/best setup is for typechecking + LSPing Django these days?
I've been leaning on pyright + django-stubs, but wondering if I'm missing something better with fewer gaps and pain points.
Hm, it doesn't seem to be dealing particularly well with imported packages that don't have type annotations. Seeing a bunch of "has no attribute" warnings. Some of the "substitute" annotations also seem to be wrong (e.g. asyncio in CPython has no annotations [in my installed version], but it's pulling some in from somewhere and they're not quite right…) It's also getting confused about lists and tuples in __slots__.
The standard library does not directly include type hints, they are stored in typeshed: https://github.com/python/typeshed
You can take a look yourself if you think some of them are wrong: https://github.com/python/typeshed/tree/main/stdlib/asyncio
The advantage is type hints can be fixed without needing to release a new version of Python. The disadvantage is there's a lot of code in the standard library that doesn't really consider how to represent it with type hints, and it can be really tricky and not always possible.
I'm surprised to see so many people moving to pyrefly, ty, and zuban so quickly. I was going to wait until some time in 2026 to see which has matured to the point I find useful, I guess some users really find existing solutions actually unworkable.
> You can take a look yourself if you think some of them are wrong: https://github.com/python/typeshed/tree/main/stdlib/asyncio
Hmm. Presumably mypy and pyrefly use the same ones, but then I don't understand why pyrefly is complaining and mypy isn't:
The definition in typeshed is this: …where is it even puling "tuple[Any]" from…(tbh this is rather insignificant compared to the noise from external packages without type annotations, or with incomplete ones… pyrefly's inferences at the existence of attributes and their types are extremely conservative…)
> Hmm. Presumably mypy and pyrefly use the same ones, but then I don't understand why pyrefly is complaining and mypy isn't:
> …where is it even puling "tuple[Any]" from…
Perhaps it's a bug in pyrefly, perhaps mypy or pyrefly is able to infer something about the types that the other isn't. I would strongly suggest checking their issues page, and if not seeing a report already report it yourself.
While there is an ongoing push to more consistently document the typing spec: https://typing.python.org/. It does not actually cover all the things you can infer from type hints, and different type hint checkers have decided to take different design choices compared to mypy and will produce different errors even in the most ideal situation.
This is one of the reasons why I am waiting for these libraries to mature a little more.
> it does not actually cover what rules you can check or infer from type hints
Indeed this is the cause of maybe 30% of the warnings I'm seeing… items being added to lists or dicts in some place (or something else making it infer a container type), and pyrefly then refusing other types getting added elsewhere. The most "egregious" one I saw was:
Where it complains: Edit: now that I have posted it, this might actually be a bug in the .join type annotation… or somethingEdit: nah, it's the loop (and the LiteralString variant of .join is just the first overload listed in the type hints)… https://github.com/facebook/pyrefly/issues/1107 - this is kinda important, I don't think I can use it before this is improved :/
I assume in your example if you update the foo declaration to the following it solves the complaint:
If so this a type checking design choice: I don't know Pyrefly's philosophy here, but I assume it's guided by opinionated coding guidelines inside Meta, not what is perhaps the easiest for users to understand.Yes, annotating the type explicitly fixes it; but tbh I'd consider that type annotation "unnecessary/distracting code litter".
As far as their philosophy goes, it's an open issue they're working on, so their philosophy seems to agree this particular pattern should work :)
I agree, but as a type checker it is a subjective choice, whether to be explicit and not make assumptions or whether to infer from certain certain patterns as probably correct and not be noisy to the user. Very glad to see they plan to fix this.
Switched to pyrefly for my research project and love the speed advantage and clarity of references! Excited for new developments
I love the speed advantage of pyrefly over (based)pyright but so far it doesn't seem to highlight as much as pyright does, for example it doesn't catch unreachable code like this:
Autocomplete for modules also doesn't work for me yet: Looking forward to have a fast language server for python though, pyright is way too slow for large projects.Hey, Pyrefly developer here, thanks for trying us out! Thanks for bearing with us with these issues you're experiencing while we're still in alpha.
We're planning on adding unreachable code diagnostics soon (github issue [here](https://github.com/facebook/pyrefly/issues/1292)). These come for free with Pyright so we don't want to regress features.
I'm happy to help diagnose/fix your autocomplete issue: it should work on modules. If you want to provide details here, on [discord](https://discord.gg/Cf7mFQtW7W), or as a Github issue (github/discord preferred) we'll fix it for you + anyone else with the problem.
It's funny because in other typed languages a string can be null because of string being a reference type as opposed to a value type. I think Python type hints make more sense than C# does in this respect, which gives me a chuckle (two of my favorite languages and the untyped one makes more sense with types). That said, not sure if you already have but I would add a github issue / ticket reporting your issue to raise awareness to the devs.
If you use the nullable types compiler option in C# (which defaults to enabled for new projects) then you need to declare you string as string? for it to be nullable :)
A quibble: Python is strongly typed. It’s not un-typed in any reasonable sense of the word. It’s dynamically typed, though: type information is on objects, not on the names referring to them.
The billion dollar mistake.
All Python types are reference types. The reference to None isn't included in the str type.
There's no way to get an actual null reference, afaik. Variables always have some value, possibly None.
(not sure what happens if you set a reference to null from C - a crash, probably?)
In some other typed languages, that is. E.g. in C++, OCaml, Haskell, F#, Kotlin, and Rust (a non-exhaustive list) you can have non-nullable strings, and other objects.
It’s worth noting that even though the runtime allows nulls (i.e. None) anywhere, Python type checkers do distinguish between optional and mandatory types.
Whats the expected error in the example you gave? That x can't be none because it was received as a str?
Yes, exactly. x would have to be str | None to be reachable.
Exactly - something along the lines of "Statement is unreachable".
isn't it perfectly valid to pass None to that function ? It's not like python enforces types at runtime nor at compile time. Right ?
It's not valid from a typing perspective, but python will let you. If you want to disregard types though then none of this matters anyway and you won't get much benefit from these tools
Sure, it's valid python to do that. By that logic you could also pass an int. But the context of this post is that you're using a static type checker.
Or that x is unused?
Yeah I have a similar experience with ty.
Looks like none of these new type checkers are ready yet.
Ty has autocomplete for imports, but it's hidden behind a toggle right now. They are still working on it. They index all the modules and functions, so you can just type the function name and it will suggest the correct import and insert it.
See the recent Talk Python podcast featuring some Pyrefly team members, https://talkpython.fm/episodes/show/523/pyrefly-fast-ide-fri...
I much prefer the rigidity of pyrefly. It enforces a standard expectation across the code base, which will lead to less surprises IMO.
The weird Literal promotion makes pyrefly unusable for me, except if I'd avoid Literal types: https://github.com/facebook/pyrefly/issues/742
Though, I'm happy with basedpyright and usually disable type checking. It's great to have so many options in language servers for Python now that pylance is locked to vscode.
Python is starting to feel a bit like JavaScript circa 2014. Remember grunt, gulp, webpack, coffeescript, babel? Now we've got pyright, mypy, pyrefly, black, ruff, ty, flake8, poetry, uv...
I used to find this kind of tooling explosion exhausting (and back then with JS it truly was), but generally speaking it's a good sign that the community is hungry to push things forward. Choice is good, as long as we keep the Unix spirit alive: small, composable tools that play nicely together and can be interchanged.
Interestingly besides typescript, javascript in 2025 is still super fragmeneted but by a bunch of well-polished tools that all do approximately the same thing. esbuild/vite(rollup)/trubopack(swc), prettier/biome/oxc, npm/bun/pnpm/yarn, bun/node/deno/worker-runtimes
For me it is easy, Spring/Quarkus/ASP.NET, if it has to be a JavaScript framework, Next.js the way Vercel intended.
Anything whatever the FE team feels like using, and the less I know about it, the better.
Next.js has been called infuriating here recently: https://news.ycombinator.com/item?id=45099922
It's just people refusing to use new tools. Prettier is so much slower than any Rust formatter, it's painful. `yarn install` takes forever. Why not switch to Bun? It's 5 minutes.
I interviewed at one company a few months ago and they said they don't use TypeScript because the compiler gets in their way, jesus chirst.
The fact that you even didn't mention webpack, once a champion, is especially sad.
Doesn't Next.js still use webpack?
A lot of the tools you list here are composable, mypy checks type hints, black formats code, because of the purpose and design ethos of those two tools they would never think to merge.
So which is it that you want, to just reach for one tool or have tools that have specific design goals and then have to reach for many tools in a full workflow?
FWIW Astral's long term goal appears to be to just reach for one tool, hence why you can now do "uv format".
"There can be only one" - just need to wait till the deathmatches are over.
Pity that is all looking like Rewrite in Rust looking for solution, instead of actually improving JIT tooling capabilities.
There is a reason us old timers mostly wait on the sidelines until the dust settles.
We have seen this movie already too many times.
"grunt, gulp, webpack, coffeescript, babel" --- except no one uses these anymore and they are dead outside of legacy software.
The problem with the python tooling is no one can get it right. There aren't clear winners for a lot of the tooling.
I think that's the point. Every now and then a language will have a small explosion of new tooling, and all you can really do is wait for it to blow over and see what tools people adopted afterwards, it feels like Python is going through a period like that at the moment.
This looks interesting. But I decided to go with ty in my projects, for the incremental approach.
How well does ty work these days? Last time I checked it still produced a lot of false positives and false negatives. (Which was very much expected, given its alpha/WIP status.)
Honestly i barely use it, just started to give it a try by adding types where it makes most sense and running with pre-commit. No issues so far. But, again, I'm a very light user.
I've tried pyrefly, ty, pyright, and basedpyright, with a large complex code base written using PyCharm, and _none_ of them do as good a job as PyCharm, particularly in discovering more complex type inheritances. It's a pity because in other respects Zed (which relies on these) is a worthy competitor to PyCharm (and much faster!) -- but the endless squiggly lines because pyrefly can't figure out the type, is annoying (and turning off the warnings is unhelpful). Hopefully one of these will get up to PyCharm's level (my money would be on ty as Astral is kicking a* these days).
PyCharm has very basic type checking, though. It's not strict.
> pyrefly, ty, pyright, and basedpyright
All of them will complain 2-4x more about your code than PyCharm. I had more than 300 typing errors when I first opened my 20k LOC project in pyright that I wrote in Pycharm.
PyCharm works great when your code is not annotated. It infers types very well. But it won't complain in a lot of cases when your code is annotated.
Related reddit post https://old.reddit.com/r/Python/comments/1ajnikt/to_pycharm_...
I think the 2nd best IDE for Python for me is Visual Studio proper, not Code. I know it sounds crazy, but Microsoft actually put some effort into their Python integration. Again, its a 2nd best, but that still says a lot about everyone else's editors.
Nowadays I'm finding myself using Zed a lot more, so maybe the story will be that all these nice Rust based tools become baked in giving it super powers for Python.
I don't think these are designed to "discovering" complex type inheritances.
They are designed for code which are more or less fully typed, as opposed to PyCharm which cobbles together a bunch of heuristics to try to make sense out of untyped code. An admirable quest, but not one I'm personally interested in.
And their insistence on only supporting this approach drove my entire team away from using PyCharm.
(From shallowly observing notifications on the 20+ typehint related issues I'm subscribed to, they seem to have kinda turned around and working toward fully supporting the python type system finally - possibly by integrating with one of the third-party type-checkers)
Another one I recently discovered and that has a very active maintainer is "zuban" or zuban-ls. It has replaced jedi-language-server for me, and aims to replace mypy as well.
Yeah, there are now 3 competitors and they all written in Rust:
- zuban
- ty (from ruff team)
- pyrefly
One year ago, we had none of them, only slow options.
Speed is one of those "Quantity has a quality all its own" things. We use very fast tools in a qualitatively different way, even though all that changed was how long it takes in seconds.
It is interesting that nobody was writing these tools in C or in C++. There are obvious ergonomic reasons, but perhaps also it matters that Rust cares a lot more about types than either of those languages.
> It is interesting that nobody was writing these tools in C or in C++
This one's easier to explain. People interested in tooling for a specific language probably want to write that tooling in that language (hence pip, poetry, mypy, jedi, etc). Normally that would be the end if it, if Python wasn't 10-100x slower than a natively compiled language. And going from Python to Rust is an order of magnitude easier than going from Python to C or Python to C++, because the compiler is so good at identifying silly mistakes. Rust is just a friendlier language.
It's just very hard to write such systems in C/C++. Even all these Rust versions are segfaulting and panicking quite a lot. So many corner and edge cases that can be found in Python code, and the memory handling is also hard.
The author of Zuban started writing it back in 2020 or 2021, so it took him more than 4 years to complete it. And he is the author of Jedi, so he had prior experience already.
I get why they'd panic, but why have enough segfaults that you noticed? So, I went and read the code.
Zuban seems to have a bunch of scary "I'm not sure if this is correct" unsafe blocks, which to me would be a red flag. I mean, it's better that there's a comment expressing the doubt, but my experience is that if you're not sure whether it's correct, it's probably not correct.
While acknowledging the risk of causing a pile-on (which I don't want!), would it be possible to have a link to them or a description of what the unsafe blocks accomplish? I'm intrigued if they are for performance or API ergonomics, if they are due to limitations of the borrow checker, the stdlib or crate dependencies.
For anyone reaching for unsafe, there are in many cases either an existing API (split_at_mut comes to mind). For others, using zero-copy or bytemuck instead of unsafe is a good idea too.
None of that is to say "never write unsafe", unsafe existing is pretty much one of the reasons for Rust to be :)
Basedpyright is not rust but it's a fork of pyright with added features that are otherwise locked in vscode
It's written in Typescript, which is a super weird choice.
Wasn’t pyright made specifically for VSCode? That would explain TS.
Yes, but why they did not write it in a compiled language? Pyright is pretty slow in large code bases and takes a lot of RAM. Javascript can be faster than python in some cases, but Python is so easily extendable with C,C++, Cython, Rust. They could use Python with one of the compiled language.
[dead]
I'm sorry, I can't take seriously any piece of software which decided to prefix the previous version's name with "based". I'm aware this is a me problem.
Definitely this. I commend author of BPyright, but clown (?) avatar, unknown identity of maintainer, and name of the fork rub me off wrong way.
Hah. I love the name. It implies that whatever the original “pyright” was doing wasn’t keepin’ it real. This new version, it’s “based” so it must be somehow more “real” and “grounded” and “legit”.
All I know is it is much more strict about stuff than pylance was.
Also a me problem!
It's also horrible for fasle positives unless your project happens to be the exact same setup as the maintainers' - I had to turn off the actual type checking on it. I've since moved wholesale to the Ty alpha and it feels a hell of a lot smarter.
It also inherits the unfortunate attitude of Pyright that it will warn against idiomatic Python (EAFP) in favour of non-idiomatic Python (LBYL):
https://github.com/microsoft/pyright/issues/1739
https://docs.python.org/3/glossary.html#term-EAFP
https://docs.python.org/3/glossary.html#term-LBYL
Sometimes dynamic Python idioms are incompatible with typed Python. I personally think that's fine, since I consider static typing a significant improvement overall.
This isn’t. They actually fixed that bug. Then they changed their minds and backed the fix back out again because they don’t think you should write Python that way:
> I think EAFP is a very unfortunate and ill-advised practice.
They want you to not write the idiomatic Python:
…and instead write the non-idiomatic version: If this were a linter then I would accept that it is going to be opinionated. But this is not a linter, it’s a type checker. Their opinions about EAFP are irrelevant. That’s idiomatic Python.Eww, what? I hadn’t seen that before. Yikes, I hope the situation’s improved. I’d be butting into that continually.
it'll be like python package managers and js web frameworks.... a new one every quarter
uv works so well for the vast majority of scenarios that I don't really see a demand for further innovation in the Python package manager domain.
yeah we all heard that story every 3 months with all the previous package managers. until there's adoption by an overwhelming majority of projects, it isn't really settled yet.
I’ve literally never heard that much buzz and excitement about Python tool before. And I’ve seen them all.
All of them had some big issue that prevented it from getting mainstream. Either it was slow, or didn’t work with existing workflow, or had complex configuration, or something that prevented gradual adoption.
uv is universally praised as the second coming Christ in Python world (and for a good reason). So no, I doubt there will be something else. Not only you need to be better than uv, you also need to have community momentum.
it still needs to add handling for binaries (the one thing conda can do that uv can't)
Sadly, the question with both this and ty is: Does it support pydantic? If not, then it's not really helpful for many people, and right now AFAIK neither supports pydantic.
Pydantic is probably the problem here, but it is what it is.
Pyrefly actually does have (experimental) support for Pydantic! https://pyrefly.org/en/docs/pydantic/
Nice, did not know.
Python's type system is the problem here, which cannot support even the native dataclass pattern and needs custom plugins written for each type checker.
The type system’s alright. It just gets especially tricky when you’re trying to check code which won’t exist until you run it. For instance, suppose you wanted to load your own module from a database with something like
It can’t know what you’re going to load until it actually does it.Things which lean heavily into metaprogramming, typically ORMs or things like Pydantic, fall into that category. I can’t hold that against the type system.
I don't think this is really true, I think native data classes can do just about everything you need, it's just that you need some tooling around them to get the same as Pydantic offers. I'm pretty sure we are going to see better native dataclass support in Pydantic going forward.
Already there is some support in Pydantic for native dataclasses: https://docs.pydantic.dev/latest/concepts/dataclasses/
I have used native dataclasses with pydantic starting with v1.5 and it’s only improved with the the newer 2.x releases.
Aside from basic inheritance and complex nested types, the pydantic ‘TypeAdapter’ is awesome for simply validating native dataclasses. It’s a little slow, but everything pydantic is =)
[dead]