Beatra All-in-One Creation

Problem Solved

Multimedia content creation—such as for images, videos, music, and speech—often requires multiple separate tools, leading to fragmented workflows, opaque costs, and complex operations. For instance, creating a voiceover video might involve separate image generation, speech synthesis, and video editing tools, with manual coordination of formats and timelines, plus managing subscriptions and task queues. Beatra All-in-One Creation aims to transform such cross-media requirements into a unified, verifiable workflow, reducing integration pain points and providing transparent payment control.

How It Works

The skill executes all Beatra operations via the bundled mcp_client.py client, avoiding REST/OpenAPI fallbacks. Core capabilities include:
- Cross-media generation: Supports images (beatra.images.generate, beatra.images.transform), videos (e.g., beatra.videos.generate, beatra.videos.animate for digital human driving), music (beatra.music.generate), and speech (beatra.speech.synthesize, beatra.voices.clone for voice cloning).
- Key steps:
1. Connection and authorization: First-time connection requires browser authorization to generate a Device Token, with automatic registration; subsequent cross-media switches need no re-authorization.
2. Minimal path selection: Calls appropriate tools based on requests; e.g., before video generation, uses beatra.models.list to check capability (like text_to_video), verifying admission duration and resolution.
3. Payment boundary management: Distinguishes free operations (e.g., beatra.wallet.get balance queries, beatra.social.tools.search social searches) from paid tasks (e.g., beatra.videos.generate); videos and cloning require an admission card with estimated cost and top-up link, user confirmation before submission.
4. Task tracking and recovery: Uses client_request_id to identify tasks, polls via beatra.tasks.get; if IDs are lost, employs beatra.tasks.list to compare input payloads (e.g., operation_key, schema_hash) for recovery.
5. Secure upload: Local media handled via beatra.assets.upload, adhering to 100 MB limits and voice cloning authorization requirements (consent_attested: true).

Boundaries and Considerations

• Payments and authorization: Generation operations consume credits; video prompt enhancement (beatra.videos.enhance_prompt) is post-paid, while other video generations require real-time estimates and user confirmation. Voice cloning needs explicit owner consent.
• Model dependencies: Features rely on model support (e.g., interface cards from beatra.models.list); incompatibility should be explained rather than substituted.
• Error handling: Preserves error.code, distinguishing task failures (e.g., insufficient_balance) from model errors, and avoids resubmission.
• Updates and uninstallation: The client auto-updates without interrupting tasks; uninstall follows the process for ~beatra connection state, avoiding direct deletion.