Releases

Core scaffolding for running LLM-powered agentic loops

Language Server Protocol implementation for Harn

High-level wgpu integration for Syphon

Linter for the Harn programming language

Opt-in code-intelligence and deterministic-tool host builtins for the Harn VM

Declarative structural rule engine for Harn — rule model, pattern compiler, and matcher built on the harn-hostlib tree-sitter machinery.

Agent-first issue tracker (SQLite + JSONL)

Downloadable on-device prompt-injection classifier for Harn (Layer 2)

Debug Adapter Protocol implementation for Harn

Typed VM definitions — Darwin-hosted Linux guests for tatara-os, authored in tatara-lisp, emitted as vfkit/qemu config

Guest runtime orchestrator — 'hospedeiro' = 'host' in Portuguese. Polymorphic supervisor dispatching GuestSpec → HVF/VZ/WASM backends. Phase H.6.

Arena-enclosed Lisp virtual environments — bumpalo-backed, sealable, serializable compilers you can author in Lisp and deploy as typed artifacts

A Rust-native programmable browser runtime built on Servo and SpiderMonkey

NixOS-equivalent — typed SystemConfig deriving TataraDomain, activation synthesizer, Linux-kernel + bootloader + services + users + filesystems, composed via tatara-nix module fixpoint

tatara IS PID 1 — typed service supervisor authored in tatara-lisp. Drop-in replacement for systemd/s6/openrc inside a tatara-os Linux guest.

Package set abstraction + nixpkgs bridge + tend-driven tatara-lisp mirror — the tatara equivalent of nixpkgs

Optimized in-memory Lisp interpreter for tatara — evaluates Sexp to Value, produces typed Derivations, composes with tatara-nix realizers

SpiderMonkey-backed API compatibility layer replacing bun_jsc

SpiderMonkey engine wrapper for Bao runtime

An AI-native version control system

Heddle's CLI verb catalog, schemas, help, and recovery contract.

An in-process property graph database management system built for query speed and scalability

OpenAI-compatible HTTP serving for the memra CUDA inference engine - single-GPU multi-model step-interleave scheduling on RTX 50-series

A component of the servo web-engine.

From-scratch CUDA LLM inference engine for NVIDIA RTX 50-series (sm_120a) and Hopper (sm_90a) - custom kernels, no frameworks

Structure-aware query CLI for the Quarb engine

A component of the servo web-engine.

功能完整的 A 股数据获取库，完全对标 pytdx

Guest runtime orchestrator — 'hospedeiro' = 'host' in Portuguese. Polymorphic supervisor dispatching GuestSpec → HVF/VZ/WASM backends. Phase H.6.

Typed VM definitions — Darwin-hosted Linux guests for tatara-os, authored in tatara-lisp, emitted as vfkit/qemu config