Report Abuse

Basic Information

Mochi is a small, statically typed programming language and toolchain designed for developers who want a compact, embeddable language for building tools, processing real-time data, and powering intelligent agents. The repository provides the mochi CLI and built-in bytecode VM for running Mochi programs, a language server and a VS Code extension for editor integration, example programs, and support for running as an MCP server to connect with agent host environments like Claude or VS Code Agent Mode. It includes prebuilt binaries, Docker images, and instructions to build from source, so users can run, test, build, serve and extend Mochi on their platform. The README emphasizes agent-friendliness, portability as a single zero-dependency binary, and testability via built-in test and expect blocks.

Links

Categorization

App Details

Features
Mochi offers a declarative, functional syntax with static types, an integrated bytecode VM optimized with constant folding and dead code elimination, and a concise CLI exposing commands such as run, test, build, serve, repl and llm. The language includes dataset queries with SQL-like clauses and joins, streams and agent declarations for event-driven code, generate blocks to invoke LLMs and produce structured outputs or embeddings, and a foreign function interface to call extern libraries. Tooling includes a Language Server Protocol implementation, a VS Code extension, Docker images, MCP server compatibility and small MCP tools like mochi_eval. The project is distributed as a single binary for portability and provides examples and test suites to get started.
Use Cases
Mochi helps developers prototype and ship agent-oriented programs quickly by combining a compact, embeddable language with runtime and editor tooling. It reduces setup friction with a zero-dependency binary and optional Docker usage, enables reproducible testing with built-in test blocks, and supports direct LLM integration through generate blocks and model configuration. The MCP server and language server let Mochi run inside agent host environments and editors for interactive agent mode, while FFI and dataset query features let developers connect to external code and structured data. Examples, a cheatsheet, benchmarks and contribution guidance make it easier to learn, extend and integrate Mochi into pipelines that need safe, testable, event-driven or AI-enabled behavior.

Please fill the required fields*