Glade Studio

We build worlds that surprise, challenge, and evolve - and the tools to power them.

Headquartered in San Francisco, CA, our small team blends technical rigor with creative ambition. We believe the future of games is living worlds: environments, stories, and relationships that grow with the player-made possible by on-device AI.

🔗 Website: gladecore.com

GladeCore - Local AI NPCs for Unreal Engine

GladeCore is our on-device system for advanced, AI-driven NPC interactions in Unreal Engine. It lets players hold dynamic, unscripted conversations with NPCs via text or voice - fully offline, with zero per-interaction cost.

Highlights

• LLM-Powered Dialogue: Local inference for natural NPC responses
• Zero Cloud Dependency: Offline by default; scale without variable costs
• Speech-to-Text (STT): Talk to NPCs via microphone
• Text-to-Speech (TTS): NPCs speak responses (local TTS or external providers)
• Data-Driven Personalities: Backstories, traits, voices via UE Data Assets—no code required
• Multiplayer Ready: Pro & Enterprise tiers support networked play

Why GladeCore?
We built GladeCore while shipping our own AI RPG and ran straight into cloud cost spikes, latency, and production complexity. GladeCore is the developer-friendly alternative: performant, affordable, and game-ready.

GladeCore NPCs aren’t “just chatbots.” They can express emotions, develop unique behaviors, remember context, and generate creative, human-like responses - all constrained and directed by your game’s data.

What you’ll find on our HF org

• Models & Adapters: Lightweight local models, LoRAs, and prompt/response assets tuned for game dialogue, emotional control, and persona consistency.
• Datasets: Clean, game-focused conversation and trait datasets designed for controllable NPC behavior.
• Spaces (Demos): Browser-based prototypes that preview our pipelines and show how personality controls affect outputs.

We ship assets that optimize for on-device use in games: small, fast, and controllable.