Moonbit-ic-cdk – A MoonBit-powered Canister Development Kit for ICP

Community Showcase
1

logo
logo1024×1024 33.4 KB

:rocket: Say hello to moonbit-ic-cdk — The first MoonBit-powered Canister Development Kit for building on the Internet Computer (ICP)

:crescent_moon: Why MoonBit + ICP?

WebAssembly (WASM) is at the core of canister smart contracts on ICP. But here’s the catch:

Most non-Rust CDKs on ICP struggle with WASM compatibility, performance bottlenecks, or simply developer complexity.

MoonBit changes that.

It’s a purpose-built, all-in-one programming language toolchain designed for WebAssembly from the ground up.

With MoonBit, you get:

  • :package: Significantly smaller WASM binaries than most toolchains
  • :high_voltage: Faster compile times and runtime performance
  • :writing_hand: A simple, pragmatic, and data-oriented syntax
  • :light_bulb: Easier to learn than Rust (but still powerful enough to build robust dapps)

That’s exactly why we built moonbit-ic-cdk — to make building on ICP faster, lighter, and more accessible.

:hammer_and_wrench: What’s inside

moonbit-ic-cdk

Module Description
ic-types Core IC types like Nat, Principal, Result
ic-principal MoonBit-native Principal implementation and utilities
ic0 Wasmtime FFI bindings to the IC system API
ic0-api-warp Higher-level abstraction over ic0, inspired by Rust’s ic-cdk
example A working canister, fully written in MoonBit and deployed on-chain

:motorway: Roadmap

:hourglass_not_done: Next 1–3 Months (High Priority)

The top development focus is completing candid-moonbit, an essential part of the ic-moonbit-cdk ecosystem:

  • LEB128 encode/decode
    Develop leb128-moonbit to handle Candid-compliant integer serialization.

  • Basic Candid types mapping and se/de implementation
    Map Candid primitive types to MoonBit types and implement corresponding serialization/deserialization.

  • Serialization/deserialization for core types
    Implement se/de for types like number, principal, and result.

  • Candid IDL Builder
    Build type serializer and value serializer generators for Candid interface description.

  • Encapsulation of http_request, http_response, and storage APIs
    Simplify building web service canisters with high-level HTTP and storage interfaces.

:crystal_ball: Future Feature Plans (Post V2)

  • Further enhancements to ic-types-moonbit
  • Async support for canister-to-canister calls
    Implement asynchronous messaging if MoonBit introduces async/await primitives.
  • Management Canister API support
  • Fully-featured CDK API
    Develop high-level abstractions for stable memory, certified data, traps, and cycles, etc.
  • JSON serialization as default

:handshake: We’re collaborating directly with the MoonBit core team to bring WASM-native development into the heart of ICP.

:books: Resources

14 Likes
2

I haven’t heard of moonbit before
Where is this language used right now and how would it compare to Motoko

1 Like
3

Hi @Gekctek

Moonbit is a very new language designed to embrace WebAssembly (Wasm). Its core development team is based in Hong Kong, and they’ve previously held several meetups in places like Singapore. They also showcased the language at last year’s Rust China conference.Previously, the Moonbit core team had a brief conversation with DFINITY as well.

Currently, Moonbit is gaining significant popularity in Asia, especially among teams working with Wasm. Many of these teams are choosing Moonbit to build their components, particularly for implementing cryptographic algorithms that can be called from JavaScript.

Motoko is an excellent language—I started using it back in 2021 and have built many DApps with it. However, compared to Motoko, Moonbit has broader recognition in the Asian region. The Moonbit team is also heavily focused on integrating AI into the compiler, with the goal of enhancing developer productivity.
Their approach includes:

  • AI-friendly language design, featuring safety mechanisms, easy testing, and reduced hallucinations from LLMs.
  • Built-in AI assistant to automate testing and documentation, streamlining development workflows.
  • Intelligent code generation with high accuracy and fast response times.

I hope to bring these developers into the ICP ecosystem through the moonbit-cdk, so that ICP can be adopted in more non-crypto industries.

3 Likes
4

Maybe worth adding this to

3 Likes
5
2 Likes
6

Hi, I’ve updated the upcoming Roadmap for ic-moonbit-cdk:

:motorway: Roadmap

:hourglass_not_done: Next 1–3 Months (High Priority)

The top development focus is completing candid-moonbit, an essential part of the ic-moonbit-cdk ecosystem:

  • LEB128 encode/decode
    Develop leb128-moonbit to handle Candid-compliant integer serialization.

  • Basic Candid types mapping and se/de implementation
    Map Candid primitive types to MoonBit types and implement corresponding serialization/deserialization.

  • Serialization/deserialization for core types
    Implement se/de for types like number, principal, and result.

  • Candid IDL Builder
    Build type serializer and value serializer generators for Candid interface description.

  • Encapsulation of http_request, http_response, and storage APIs
    Simplify building web service canisters with high-level HTTP and storage interfaces.

:crystal_ball: Future Feature Plans (Post V2)

  • Further enhancements to ic-types-moonbit
  • Async support for canister-to-canister calls
    Implement asynchronous messaging if MoonBit introduces async/await primitives.
  • Management Canister API support
  • Fully-featured CDK API
    Develop high-level abstractions for stable memory, certified data, traps, and cycles, etc.
  • JSON serialization as default
1 Like
7

Edit: I see it now under Future plans point 3.

What’s the general idea behind a Candid implementation?

  • With Motoko the types basically map one to one.
  • With the Rust SDK there’s automatic mapping to Rust types.
  • With the Azle SDK you need to annotate methods with Candid types.

Does the MoonBit language support automatic mapping approaches like the Rust SDK or is an approach like the Azle SDK more straightforward in MoonBit?

8

Hi @sea-snake , I’m glad you’re interested in this topic!

At this stage, it seems that MoonBit can adopt a similar approach to the Rust CDK. While developing the V1 version, I experimented with serialization/deserialization (de/se) of some basic types and found that handling the raw bytes correctly was often sufficient — although it’s necessary to include some additional headers, such as those defined by DIDL (e.g. length prefixes and type descriptors).

However, my current focus is on supporting JSON serialization and deserialization, since MoonBit has relatively good native support for JSON. I believe the first wave of MoonBit CDK users will also be more familiar with JSON-based workflows.

In V2, I plan to prioritize implementation of point 1 and point 2 from the roadmap. Support for Candid will be introduced in V3. Here’s the planned progression:

  • V1: Developers need to interact with low-level APIs and handle serialization manually in many cases.
  • V2: Developers will be able to use a higher-level API similar to the Rust CDK, and deploy services such as web server–like frontend hosting directly on IC.
  • V3: Focus on deeper compatibility with IC blockchain features, such as threshold signatures and certified data (certification).

If you have other suggestions, feel free to reach out — I’m still planning things carefully. Thanks for your interest and support!

2 Likes
9

Just a quick update:
After some research and discussions over the past period, I have adjusted the implementation roadmap for ic-moonbit-cdk.
The new roadmap has now been updated.

1 Like
10

Update:

Look what I’ve discovered—Juno! This is an awesome project.

It features a beautiful, minimalist UI and comprehensive documentation and API. I actually first heard about it through its SNS, and I regret not having started exploring it sooner.

I’m thinking that integrating Juno will be a major goal for me after finishing candid-moonbit(this will officially establish Moonbit as a tool for developing canisters). I believe it’ll be an excellent tool for deploying canisters and providing web services.

It’s a somewhat long-term goal (I’ve been quite busy lately), but I’m really excited about it!

@peterparker

2 Likes
11

So great to hear! Likewise, that’s super exciting :rocket:

1 Like