Releases

Core library for tree-sitter language pack - provides compiled parsers for 306 languages

The world's first multi-agent shell with typed functional pipelines and multi-modal AI

Opinionated Arora runtime: an engine pre-wired with the native behavior-tree control nodes and the Semio backend.

The Arora engine: loads and executes Arora modules (wasm and native hosts).

Arc framework runtime: Actix/Tera web machinery, event-sourcing wiring, and the application builder seam.

Rust-centric multi-language HTTP framework with polyglot bindings

Chip-selecting radio facade for HiSilicon embedded Rust

A Rust-based terminal coding agent with modular architecture supporting multiple LLM providers

ACP bridge and client implementation for VT Code

The simple image manipulation program

Core library for VT Code - a Rust-based terminal coding agent

The world's first multi-agent shell with typed functional pipelines and multi-modal AI

SecAPI bindings

FFI bindings to SecAPI

A low-level windowing system geared towards making audio plugin UIs

Chip-selecting radio facade for HiSilicon embedded Rust

Model Context Protocol (MCP) client, connection pooling, and tool discovery for VT Code

Unified UI crate for VT Code: design system, theme registry, and TUI framework

LLM provider abstraction, client implementations, and streaming for VT Code

Skill types, discovery, loading, and validation for VT Code

Build scripts for compiling librnp from source

Cross-platform shell execution helpers extracted from VT Code

A queryable code knowledge graph tool built on LadybugDB and tree-sitter

The world's first multi-agent shell with typed functional pipelines and multi-modal AI

Config loader components shared across VT Code and downstream adopters

PMAT - Zero-config AI context generation and code quality toolkit (CLI, MCP, HTTP)

Chip-selecting radio facade for HiSilicon embedded Rust

A static analysis platform for Rust program analysis and verification

The world's first multi-agent shell with typed functional pipelines and multi-modal AI

Evaluate programs, CLI commands, programming languages, AI frameworks, and VM/sandbox systems for agentic AI use across four axes — token efficiency, determinism, reliability, and safety — under popular tokenizers (OpenAI GPT-4/GPT-4o, Anthropic Claude). Includes a CLI effect classifier, curated language/framework/VM profiles, and a self-describing ontology.