Tryll

What Tryll does

Tryll Engine provides on-device AI middleware for game developers, enabling AI to run locally on the player's hardware. This eliminates cloud costs, data exposure, and third-party dependencies. Game studios use Tryll to integrate AI into gameplay in novel ways, offering a foundation for developers to build upon without re-engineering their existing pipelines. The platform is designed to leverage the GPU in every player's machine, offering a new medium for game design.

Products

• Tryll EngineOn-device AI middleware for games, providing Unreal & Unity plugins, game state awareness, context-aware AI, custom AI characters & companions, voice integration (TTS/STT), AI agents, and dynamic content generation.
• Tryll PluginA simple API for prompts, agents, tools, and actions that integrates with Unreal and Unity engines.
• Tryll RuntimeRuns on the player's GPU (NVIDIA, AMD, Intel auto-detected) to process AI models locally.
• Tryll SDKSoftware Development Kit for integrating Tryll Engine, currently available for Unreal Engine with Unity planned. Supports LLM, STT, TTS, VLM, and Agents.
• Tryll AssistantA free AI overlay that runs on any game to demonstrate local LLM inference on player hardware without integration.

Key capabilities

• On-device AI processing
• No cloud costs
• No data exposure
• No third-party dependencies
• Runs on player GPU (NVIDIA, AMD, Intel auto-detected)
• Low latency (<50ms)
• Offline functionality
• Unreal Engine plugin
• Unity plugin
• Simple API for integration
• LLM support
• STT (Speech-to-Text) support
• TTS (Text-to-Speech) support
• VLM (Vision-Language Model) support
• AI Agents
• Engine-agnostic (for custom builds)
• Game state awareness
• Context-aware AI
• Custom AI characters & companions
• Dynamic content generation

Use cases

• Knowledge Systems: Connect game documentation and lore to AI for real-time player questions, reducing support load and eliminating alt-tabbing.
• Persistent Memory & Relationships: NPCs that remember conversations, form opinions, and evolve over time, enabling deep companion systems without extensive dialogue branching.
• Agent Actions: AI triggering game mechanics like spawning enemies, modifying environments, and distributing rewards for dynamic gameplay.
• Runtime Content Generation: Generating quests, lore, and world events based on player context for infinite content without static databases.

AI approach

Tryll Engine provides on-device AI middleware for game developers, running AI locally on the player's hardware using their GPU. This approach eliminates cloud costs, data exposure, and third-party dependencies. It enables features like knowledge systems, persistent memory for NPCs, agent actions, and runtime content generation.

Tech named: Local Language Models, LLM, STT (Speech-to-Text), TTS (Text-to-Speech), VLM (Visual Language Model), Agents, GPU (NVIDIA, AMD, Intel)

Industries served

• Computer Games

What it says sets it apart

• Runs 100% on player hardware, eliminating cloud costs and making tokens free.
• Personal data never leaves the device, addressing GDPR, AI Act, and export law compliance concerns.
• Everything works fully offline, removing third-party API dependencies and ensuring uptime.
• Scalable pricing model: start free, pay when the game succeeds, based on copies sold.
• Built by veterans of game development and applied AI.
• Supports both Unreal Engine and Unity.
• Low latency AI processing (<50ms).

Funding rounds we track

From the AI funding tracker — rounds as reported by the linked publications.