An unofficial Rust library for the OpenAI API.
Warning
Currently in alpha, not yet stable enough to be used in any production applications. See Implementation Progress.
- Instead of accessing all functions as methods on a single client-like structure, functions should be accessed from their own modules.
- Environmental variables should be the prioitized method of authentication, but you shouldn't be forced to do things this way. (Currently, you kind of are, because the API key is needed during compile-time.)
- This is a LIBRARY, not a WRAPPER! The goal here isn't to just give some basic wrapper functions for making HTTP requests, it's to "rust-ify" things. We want to create the illusion that the OpenAI API was made in Rust first!
- What is this, C? No, it's Rust! We follow the object-oriented paradigm, not the procedural one.
What this mainly means is less
create_completion()
, moreCompletion::new()
.
This library is not yet on crates.io unfortunately, but you can use a GitHub release as a dependency like so:
[dependencies]
openai = { git = "https://github.com/valentinegb/openai", tag = "VERSION" }
or if you clone the repository:
[dependencies]
openai = { path = "openai", version = "VERSION" }
where VERSION is the version or release you want to use,
for example v1.0.0-alpha.1
with git and 1.0.0-alpha.1
locally.
If you use git, you can also use the main
branch of this repository (branch = "main"
),
but I don't recommend it, as I can't guarantee every commit is error free.
I'm still working on making examples in the examples
directory.
If you're looking to work with the completions
module, you're in luck!
Because that's the only module there is an example for right now.
For other modules, you can look at the tests
submodules for some reference.
Examples come slowly because this project, in its current state, changes very quickly, and it's not fun making sure all examples accurately reflect the latest version. But, they are coming, don't worry!
An error you will likely run into, and a hopefully pretty self-explanatory one.
For the library to even build, you must have an environment variabled named OPENAI_KEY
which is set to,
you guessed it, your OpenAI API key. Without your API key, this library can't do anything.
In fact, the library won't build without it because at compile-time it uses your key to fetch all available models
and generate the ModelID
enumerator.
In a development environment,
this can be best resolved by created a .env
file in the root of your project with the following contents:
OPENAI_KEY=put-your-api-key-here
Then, you need to load the contents of your .env
file when your program starts.
For this, I recommend a crate such as dotenvy.
██████████
Models
████████░░
Completions
██████████
Edits
░░░░░░░░░░
Images
█████████░
Embeddings
░░░░░░░░░░
Files
░░░░░░░░░░
Fine-tunes
░░░░░░░░░░
Moderations