OpenAI builds and deploys advanced AI models like GPT-4o for autonomous agents and workflows.
Quokka
FreeStarting price
0Popularity
About Quokka
Quokka is a strictly self-hosted programming language designed for clarity and reliability in AI workflows. It emphasizes deterministic behavior through explicit error handling with Result and Option types, avoiding hidden control flow for exceptions. Variables are immutable by default, with mutability requiring explicit opt-in, which improves code traceability. The language includes a multi-stage bootstrap process that cryptographically verifies correctness, ensuring the compiler and toolchain are built in Quokka itself. Quokka supports seamless LoRA fine-tuning via Unsloth through a typed JSON IPC boundary, enabling integration with Python and PyTorch without bloating the core language. The toolchain is simple to install, automatically configuring the system PATH, file associations, and editor setup in a single step.
Key features
- Strictly self-hosted compiler and toolchain
- Result and Option error handling
- Immutable-by-default variables with explicit mutability
- Multi-stage bootstrap with cryptographic verification
- LoRA fine-tuning via Unsloth integration
- Typed JSON IPC boundary for Python/PyTorch interoperability
- Automatic PATH and editor configuration on installation
- Built-in verifier for self-hosting correctness
Use cases
- Fine-tuning large language models with LoRA adapters
- Building reliable AI pipelines with explicit error handling
- Developing deterministic workflows for machine learning tasks
Pros
- Self-hosted compiler and toolchain written entirely in Quokka
- Deterministic behavior with explicit error handling via Result and Option types
- Built-in support for LoRA fine-tuning via Unsloth
- Immutable-by-default variables with explicit mutability opt-in
- Multi-stage bootstrap process for cryptographic verification of correctness
Cons
- No Linux builds available at the time of writing
- Project is in early development (v0.2.0)
- Requires manual PATH configuration after extraction
Frequently asked questions about Quokka
What is Quokka and what does it do?
Quokka is a strictly self-hosted, deterministic programming language designed for clarity and reliability in AI workflows. It features explicit error handling with Result and Option types, immutable-by-default variables, and a multi-stage bootstrap process that cryptographically verifies correctness.
Who is Quokka suitable for?
Quokka is suitable for developers and researchers working with AI workflows who prioritize code clarity, reliability, and deterministic behavior. It is particularly useful for those who need explicit control over mutability and error handling.
How does Quokka integrate with machine learning libraries like PyTorch?
Quokka integrates with Python and PyTorch through Joey, a standard library extension that uses a strictly typed JSON IPC boundary. This allows Quokka code to spawn isolated Python processes for tasks like LoRA fine-tuning without bloating the core language.
What platforms does Quokka support?
Quokka supports Windows, macOS (both Apple Silicon and Intel), and Linux (Ubuntu/x86_64). The toolchain installer automatically configures the system PATH, file associations, and editor setup for a streamlined setup experience.
How can I verify that Quokka is truly self-hosted?
Quokka includes a `quokka check-self` command that re-bootstraps the interpreter through multiple stages and compares cryptographic hashes of the AST and output at each stage. This process verifies that the compiler and toolchain are built in Quokka itself.
How do I get started with Quokka?
To get started, download the Quokka standalone zip archive for your platform, extract it, and add the folder to your system PATH. Then run `quokka run hello.qka` to execute your first script. Detailed guides are available for installation, error handling, and ML integration.
