Skip to content
@Artificial-Humanity

Artificial Humanity

Artificial Humanity 👤✨

Welcome to Artificial Humanity! We are a research and engineering studio dedicated to building AI projects that touch upon the space between machine intelligence and human emotion.

Our core philosophy is simple: we believe artificial intelligence should run locally, respect privacy, and engage with human feeling. No remote cloud APIs, no data locks, and no delay. Just native, high-performance engines that feel alive.


🎙️ Introducing Project Prosodia

Our flagship initiative is Project Prosodia—an open-source framework for expressive, continuous, emotion-driven speech synthesis. It splits speech generation into a protocol-based pipeline, separating semantic interpretation (the Director) from vocal performance (the Actor, powered by StyleTTS2), so digital books are read with dynamic pacing, dramatic pauses, and real-time voice blending—no robotic conveyor-belt narration here.

Fun facts: we're hunting for the exact Valence-Arousal-Tension coordinate for "the hero's tragic sacrifice," to get the Actor's voice to crack slightly. And since the engine compiles and runs entirely offline, you'd still have a perfect narrator even stranded on a mountaintop with a laptop.


🧑‍⚖️ Introducing Council of Experts

Our second studio initiative is Council of Experts—a native macOS orchestration platform that runs a configurable council of LLM experts (Claude, Gemini, GPT, Grok, and local models) in parallel, has them critique and revise each other's drafts, and synthesizes their consensus through a Chairman model. It's built on the same philosophy as Prosodia: a Rust core with a native Swift/SwiftUI FFI layer, supporting live streaming, workspace file context, and multi-turn session persistence.


⚙️ Introducing FerroStep

FerroStep: a stepped path from a hollow initial-state node to a glowing terminal-state node

Our newest initiative is FerroStep—a data-driven state-machine referee for multi-agent loops. Instead of hiding orchestration state inside a framework's graph object, FerroStep treats your own database as the single source of truth: AI agents and humans act as peers on one ledger, while a pure Rust core rules on every transition—who may move what, what a retry costs, and when a human takes over. Loop ceilings survive crashes, escalation is a routing decision rather than an error, and the whole system is inspectable with a database browser. True to the studio philosophy: an engine, not a platform—no runtime, no scheduler, no hosted anything.


Thanks for stopping by, and happy coding! 👤✨

Popular repositories Loading

  1. Prosodia Prosodia Public

    Modern, local-first cross-platform text-to-speech engine. Driven by on-device Gemma 4 directorship via LiteRT-LM and highly expressive voice synthesis pipelines powered by our proprietary Prosodia …

    Swift 1

  2. Sonora Sonora Public

    Sonora is an expressive and directable text-to-speech (TTS) actor model based on the Matcha-TTS architecture. Part of the Prosodia ecosystem, it adds custom emotion (VAT) conditioning and continuou…

    Python 1

  3. .github .github Public

    Projects that touch upon the space between machines and human emotion

  4. Council-of-Experts Council-of-Experts Public

    Swift

  5. Lucida Lucida Public

    Generate images and video with Google Gemini, Veo, Runway, Kling, a local ComfyUI, FLUX, Stability AI or OpenAI — a single-binary CLI and MCP server, in Rust

    Rust

  6. FerroStep FerroStep Public

    A lightweight, data-driven multi-agent framework designed to map complex loops directly to your database.

    Rust

Repositories

Showing 6 of 6 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…