Voxta to Backyard AI Character Conversion

Moving a character from **Voxta** to **Backyard AI** (formerly Faraday) involves down-sampling a hardware-integrated "Companion Entity" into a format optimized for linguistic roleplay and local model orchestration. Voxta is designed for high-performance voice and visual interaction, while Backyard AI is optimized for depth of narrative and model-specific prompt engineering.

1. Deconstructing the Voxta Schema

Voxta profiles are structured around a rigid, hardware-aware JSON schema. Key fields like Personality and Profile are used to drive weight-based inference systems. In the Backyard AI ecosystem, these distinctions are often flattened into a more traditional description block. Our converter identifies these specialized Voxta keys and intelligently merges them into the personality and description fields that Backyard AI's engine expects.

2. Handling UUID and Identity Logic

Every Voxta character is anchored by a specific UUID (Universally Unique Identifier). While Backyard AI supports internal identification, its import process is more flexible regarding identity metadata. Our tool preserves the Voxta identity tags within the new JSON to ensure that any cross-platform lore or world info remains consistent, while stripping out hardware-specific triggers that might confuse Backyard's text-only inference engine.

3. Dialogue Normalization for Roleplay

Voxta dialogue examples (MesExample) are often tailored for Text-to-Speech (TTS) clarity, sometimes featuring specific SSML tags or voice cues. Backyard AI, focusing on the written word, performs best with clean, narrative-driven dialogue. Our engine scans the Voxta source, removes TTS-specific artifacts, and reformats the dialogue into the standardized {{char}}: "..." format that ensures maximum compatibility with the models hosted within the Backyard application.

4. System Prompt and Instruction Mapping

Voxta and Backyard AI handle "System Instructions" differently. Voxta's instructions are often hidden or managed by its companion backend. Backyard AI gives the user more direct control over the system_prompt and post_history_instructions. Our converter ensures that the "intent" of the Voxta companion—its core constraints and behavioral goals—is explicitly mapped into the instructions block of the generated JSON, preventing the character from "drifting" during narrative-heavy scenes in Backyard.

5. The Security of Local Execution

Just like the tools we bridge, our converter prioritizes privacy. By using a "Zero-Server" architecture, we perform the entire schema translation within your browser's RAM. Your companion's intimate details, creative backstory, and unique voice settings are never transmitted across the open web. This ensures that you maintain total control over your digital assets as you move them between the industry's leading local AI platforms.