beatra-ai/beatra-skills / plugins/beatra/skills/beatra
beatra
VerifiedCreate AI images, videos, music, and voice, look up public social data, edit visual results, and keep generated assets easy to find and manage. Beatra Universal turns text into images, images and references into video, ideas into songs or instrumentals, and scripts into narration, while also supporting custom voice creation and public social media data lookup across YouTube, TikTok, Instagram, and X. Use it as an AI media generator and AI creative suite for text-to-image, image-to-video, AI video editing, AI music generation, text-to-speech, multilingual voice-over, voice cloning, multimodal content creation, and YouTube or TikTok data for posts, comments, accounts, and trends. Reuse source media, returned files, and public social findings across formats, follow production progress, and view actual credit use for completed creative work.
references/speech-and-voices.md
View on GitHub# Speech, dialogue, voices, cloning, and audio tools
Use this reference alongside [speech and audio](speech-and-audio.md) for spoken narration, Smart Voice Director preparation, native multi-speaker dialogue, voice discovery/audition/management, voice cloning, speech conversion, transcription, and sound effects.
## Voice discovery, audition, updates, and deletion
- **`beatra.voices.list`** (free): browse 866 curated presets (`category="preset"`) and account clones (`category="cloned"`, `status`: `"ready"`, `"pending_activation"`, `"expired"`, `"any"`). Filter with bilingual `search` (Chinese or English), facets (`language`, `gender`, `age_group`, `use_case`, `emotion`, `model`), or capability flags (`supports_dialogue`, `supports_speech_to_speech`, `supports_direction`, `supports_emotion`, `supports_performance_tags`, `supports_context`). Inspect all 11 returned `capabilities.*` metadata fields (`supports_direction`, `supported_emotions`, `supports_performance_tags`, `supported_dialects`, `supports_context`, `supports_previous_text`, `supports_next_text`, `supports_pronunciation_pinyin`, `supports_pronunciation_ipa`, `supports_speech_to_speech`, `supports_dialogue`) before matching a voice to synthesis, dialogue, direction, or conversion.
- **`beatra.voices.preview`** (free): fetch an audition sample (`preview_url`, `kind`) so the customer can listen before spending credits.
- **`beatra.voices.update`** (free): update a cloned voice's `display_name`, `gender`, `age_group`, `use_case`, `persona`, or `language`, or set/clear the account default voice (`make_default: true`/`false`).
- **`beatra.voices.delete`** (free, **Mandatory Confirmation #3**): two-step permanent deletion.
1. Call once with `voice_id` only (**omit `confirmation_token`**). Present the returned `voice`, `paid_slot_warning`, `irreversible_notice`, and 15-minute `confirmation_token` expiry.
2. Only after explicit user confirmation, call again with `voice_id` and `confirmation_token`. Never chain both steps automatically without waiting for confirmation.
## Prepare expressive scripts cleanly (`beatra.speech.prepare_script` & `beatra.speech.validate`)
For expressive pacing, dramatic pauses, emotion cues, or pronunciation hints, replace manual slash-counting/stripping or raw tag writing with Smart Voice Director:
1. Present the production card (live price from `beatra.models.list` `{"capability": "speech_script"}`, one paid call, `client_request_id`, stop/top-up terms) and get approval.
2. Call `beatra.speech.prepare_script` with `input`, optional `instructions`, optional target `voice`/`model`, `language` (`"auto"` default), optional `context: {"before": ..., "after": ...}`, and `client_request_id`. Smart Voice Director preserves the customer's exact spoken words while planning pauses and expressive delivery matched to that voice's capabilities.
- Maximum **2 concurrent** `prepare_script` tasks per account.
- Check returned `directed` and `scripts[]` (`synthesis`, `summary`, `voice_selection`, `alternatives`, `estimate`, `warnings`). Any entry with `synthesis: null` cannot be synthesized as-is — resolve `warnings` first.
- **48-hour same-script reuse**: to test a different voice within 48 hours on identical wording, instructions, language, and context, call `beatra.speech.prepare_script` again with the same inputs changing only `voice` (and `model` if needed) so Beatra re-compiles at zero additional charge (account balance must still cover initial estimate plus active script tasks).
3. Pass each ready `scripts[].synthesis` object unchanged plus a new `client_request_id` to `beatra.speech.synthesize` — but run free **`beatra.speech.validate`** first.
```bash
python3 scripts/mcp_client.py call beatra.speech.validate
```
```json
{
"voice": "vc-preset-warm-female",
"input": "Sunlight reaches the kitchen table before anyone else does.",
"model": "auto",
"subtitle_enable": true
}
```
Inspect `valid`, `warnings`, `weighted_characters`, `estimated_duration_seconds`, live `price_credits`, and `execution` (`sync` vs `async`).
## Synthesize single-speaker speech (`beatra.speech.synthesize`)
1. Reuse a voice ID the user selected or pick from `beatra.voices.list` / `beatra.voices.preview`.
2. Run free `beatra.speech.validate`.
3. Call `beatra.speech.synthesize` with the final `voice`, `input`, and `client_request_id`. Keep `subtitle_enable` enabled (`true` by default; never send `subtitles` on `beatra.speech.synthesize`).
```bash
python3 scripts/mcp_client.py call beatra.speech.synthesize
```
```json
{
"voice": "vc-preset-warm-female",
"input": "Sunlight reaches the kitchen table before anyone else does.",
"subtitle_enable": true,
"client_request_id": "speech-opaque-1"
}
```
Deliver the returned audio link with its duration, format, and size in plain words, and the free subtitle files (JSON/SRT/VTT) when present. Keep the artifact ID, MIME type, `speech_start_seconds`, `speech_end_seconds`, and `integrated_loudness_lufs` in your notes for the next step. Speech billing uses weighted characters (each Han ideograph counts as 2, every other character counts as 1), linear per 1,000 weighted characters without rounding up below 1,000; failed tasks are refunded.
## Native multi-speaker dialogue (`beatra.speech.synthesize_dialogue`)
Use `beatra.speech.synthesize_dialogue` (`{"capability": "text_to_dialogue"}`) when 2 or more speakers interact in one continuous scene:
- Limits: **2–50 turns**, **<= 2,000 total characters per call**, **<= 10 unique voices** with `capabilities.supports_dialogue == true`, including **<= 4 ready cloned voices** (`status="ready"`).
- Always validate first with free **`beatra.speech.validate_dialogue`** before showing the production card and submitting `beatra.speech.synthesize_dialogue`.
- Split longer scripts scene-by-scene (`previous_text` / `next_text` for continuity). If a cast needs > 10 voices, > 4 cloned voices, regional dialects, or non-dialogue voices, fall back to segmented `beatra.speech.synthesize` and assemble in audio/video finishing.
- Returns free speaker-labeled `output.subtitles` and per-turn `segments[]` timing (`start_seconds`, `end_seconds`).
```bash
python3 scripts/mcp_client.py call beatra.speech.synthesize_dialogue
```
```json
{
"turns": [
{
"voice": "vc-preset-host",
"speaker": "Host",
"input": "Welcome back to the studio."
},
{
"voice": "vc-preset-guest",
"speaker": "Guest",
"input": "Glad to be here today."
}
],
"subtitle_enable": true,
"client_request_id": "dialogue-opaque-1"
}
```
## Clone an authorized voice (`beatra.voices.clone` & `beatra.voices.activate`)
Possession of an audio file is not consent:
1. **Optional paid cleanup**: if the recording contains noise, echo, or music, confirm credits via `{"capability": "audio_voice_isolation"}` and run `beatra.audio.process` (`operation: "isolate_voice"`, 4.6 s – 5 min, <= 100 MB) on its own card before cloning.
2. **Mandatory Confirmation #1 (`consent_attested: true`)**: obtain explicit confirmation that the user owns the voice or holds the owner's authorization before uploading or calling `beatra.voices.clone`.
3. **Select the clone model** as described in [speech and audio](speech-and-audio.md) §5. The `voice_clone` cards from `beatra.models.list` carry each model's price, constraints, and limits.
4. Show the six-field production card for `beatra.voices.clone` with the live clone price before minting `client_request_id`, submit once, then poll with `beatra.tasks.get`. Handle any refusal per [billing, errors, and recovery](billing-errors-and-recovery.md).
5. **Mandatory Confirmation #2 (`beatra.voices.activate`)**: when a clone result or a listed voice has status `pending_activation`, follow §5 Step E of [speech and audio](speech-and-audio.md) before using the voice.
## Speech conversion, transcription, SFX, and audio finishing
- **Speech-to-speech (`beatra.speech.convert`)**: `{"capability": "speech_to_speech"}`, `audio` (1 s – 5 min, <= 50 MB), target `voice` (`capabilities.supports_speech_to_speech == true`), optional `remove_background_noise: true`. Own production card required.
- **Transcription (`beatra.audio.transcribe`)**: `{"capability": "speech_to_text"}`, `audio` (0.5 s – 60 min, <= 100 MiB), `diarize`, optional `num_speakers` (`1–32`), `tag_audio_events`, `subtitles: true` (JSON/SRT/VTT included). Own production card required.
- **Sound effects & loopable beds (`beatra.audio.generate_sfx`)**: `{"capability": "sound_effects"}`, `prompt`, `duration_seconds` (0.5–30.0 s), optional `loop: true`, `prompt_influence`. Own production card required.
- **Audio finishing (`beatra.audio.process`)**: paid per step (`isolate_voice`, `separate_stems`, `trim`, `loop`, `fade`, `normalize_loudness`, `convert_format` incl. `ulaw_8k_wav` telephony). See [speech and audio](speech-and-audio.md) and [music editing and audio](music-editing-and-audio.md) for duration bounds, `dimensions.stem_mode` matching, daily limits, and one-at-a-time execution rules.
```bash
python3 scripts/mcp_client.py call beatra.audio.generate_sfx
```
```json
{
"prompt": "Warm rain outside a quiet cafe window, seamless ambience",
"duration_seconds": 15,
"loop": true,
"client_request_id": "sfx-opaque-1"
}
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |