AI Voice Studio

The Core Problem It Solves

Voice production projects rarely consist of a single, one-off generation. You might need to narrate a long article, where certain sections require cloning a specific voice, while others must be dubbed into another language. This type of complex, multi-step, multi-path task—blending long-form production, multilingual work, and cloning—demands a coordinator to manage state, sequence, and cost, rather than forcing you to juggle between different tools manually.

How It Works & Core Capabilities

AI Voice Studio (@beatra-ai/beatra-ai-voice-studio) acts as a unified orchestration layer for voice production. Its operational model is as follows:

1. Smart Routing & Intent Recognition: The skill first analyzes the user's instruction. If the goal is clear and covered by another specialized skill (e.g., a short-form voiceover short-form-voiceover-audio or a podcast ai-podcast-voiceover), it routes the task out. For complex, mixed, or ambiguous requests, processing continues within this studio.
2. Production Card & Confirmation Flow: For tasks handled in-house, it prepares a precise Production Card containing:
   • The final confirmed text (or a version refined for reading and approved by you).
   • A real-time verified and accurate voice_id (confirmed as ready via beatra.voices.list).
   • A real-time model selection and a cost estimate (based on beatra_weighted_characters).
   • All control parameters (e.g., speed, volume).
     You must approve this card before any potentially costly operations proceed.
3. Workflow Orchestration & Ledger Management: For multi-segment, multilingual, or hybrid tasks, the skill maintains a production ledger that clearly records the sequence for cloning, languages, and chapters. It ensures results are delivered in the approved route and order, and when modifications occur, it only re-synthesizes the explicitly approved segments.
4. Cost Control Boundaries: The skill strictly distinguishes between free planning and paid execution. Steps like text grooming, voice listing, preview links, model comparison, and fee estimation are not billed. Costs are incurred only when calling beatra.speech.synthesize or beatra.voices.clone. It provides clear estimates and requests final confirmation before any paid operation.

Applicability & Important Notes

Note the explicit boundaries of this skill:
* It is not: A video editor, lip-sync tool, media publisher, or audio transcription/waveform editor. Its focus is text-to-speech conversion and associated voice cloning workflows.
* Core Limitations: All interactions are conducted via the bundled scripts/mcp_client.py; no other fallback methods are used. For voice cloning, explicit authorization is mandatory, and a detailed "clone entry card" is presented to confirm balance and cost before operation.
* Technical Dependencies: Voice and model compatibility (e.g., language support, model list) must be confirmed via real-time API queries; only voices with a ready status are recommended. The model: "auto" set is a candidate list calculated in real-time based on a voice's compatible_models, not a fixed option.