Join a video meeting as an AI bot with voice, avatar, and screenshare across four operating modes.
Design & media
Store PA Broadcast Pack
Try itBuild dated store PA announcement reads for one campaign window: open, promo, flash, and close. This in-store announcement pack and retail PA voice studio records store promo voice lines you can play on the shop floor through the campaign dates. Use it for store opening announcements, flash-sale PA, closing announcements, and in-store promo voiceover.
What it does
Give a dated campaign four spoken PA reads: open, promo, flash, and close. One store voice. Play them through the campaign window.
The skill document
Store PA Broadcast Pack
Give a dated campaign four spoken PA reads: open, promo, flash, and close. One store voice. Play them through the campaign window.
Scope and adjacent routes
Use this Skill when a store needs dated in-store PA promo reads for one campaign window.
Route music loops to livestream-bgm-pack. Route live-commerce scripts
to live-commerce-script-studio. Do not generate or mix music in this
pack.
Inputs and defaults
The hard inputs are:
- the store or campaign name the floor should hear;
- the campaign window dates;
- the offer facts for open, promo, flash, and close, or permission to draft those four reads from the supplied offer;
- the target language;
- a pronunciation table whenever brand names, SKUs, or place names appear.
If those names exist and the table is empty, stop and collect the
readings. Do not invent an offer or a date window. A speaker sample is
optional. Without a sample, choose a catalog voice with
beatra.voices.list.
Defaults:
- four dated reads: open, promo, flash, and close;
- one store voice for the whole window;
languageomitted unless the user asks for a language;model: "auto"only when every voice-compatible live card supports the language;format: "mp3";speed: 1.0;- one synthesis call per read so a single read can be redone.
Keep each submitted input at or below 50,000 characters. Split on
sentence boundaries, never mid-sentence. Write short spoken sentences.
Do not send music cues, ducking notes, or playlist labels into input.
Planning and casting are free. Clone, if used, and each synthesis block are paid.
Clone the store voice only when the user asks. Confirm consent first: the
user must state this is their voice or that the speaker authorized this
cloning use. Access to a file is not consent. A suitable sample is
currently about 10 to 300 seconds, no larger than 20 MiB, and clean
single-speaker speech. Only after consent, inspect the sample and upload
it through the bundled client. Show the live voice_clone admission card
before any clone client_request_id: route voice_clone, tool
beatra.voices.clone, live price for one successful activated voice,
the provisional estimate, the 600-credit gift usually cannot start this
clone, https://console.beatra.ai/topup, starter ¥29 / 11,000 credits.
Do not recommend ¥198. Do not offer a free clone. Do not create
client_request_id until the user confirms they have topped up or already
have enough credits. Submit beatra.voices.clone exactly once:
{
"sample": {"type": "artifact", "artifact_id": "artifact_..."},
"display_name": "Store Voice",
"consent_attested": true,
"model": "auto",
"client_request_id": "opaque-store-clone"
}
Poll with beatra.tasks.get until terminal, then freeze that
voice_id.
Golden path
Submit every Beatra call through the bundled scripts/mcp_client.py. Do not
configure a host Beatra Connector. Do not use REST/OpenAPI as a fallback.
- Build the campaign ledger: window dates, slot ID, spoken text, and intended filename for open, promo, flash, and close.
- Call
beatra.voices.listwhen a voice is still needed. Keep onevoice_idfor the whole window. Never put a display name invoice. - Read the live
text_to_speechcard withbeatra.models.list. Confirm the voice, the four reads, weighted-character total, current estimate, and oneclient_request_idper read before the first paid TTS. - Synthesize each read with
beatra.speech.synthesizeexactly once. Poll each speech task withbeatra.tasks.getuntil terminal. - Deliver dated labeled clips with slot IDs, plus actual duration, MIME
type, URL or artifact, resolved model, and
billing.net_charged_credits. - Review that dates, prices, and the store name match the ledger. Redo only the changed read.
Decisions that require confirmation
Confirm the voice and the campaign estimate before the first paid TTS. A changed read is one new paid block.
Recovery
Recover a lost create response only with that read's identical frozen
payload and ID. A changed text is a new ID. Recover a lost task ID
through beatra.tasks.list and beatra.tasks.get before any resend.
Cancel only at the user's request. On 409, keep polling the same task.
On insufficient_balance for clone or TTS, keep
https://console.beatra.ai/topup exact and retry the same frozen ID only
after the user says they have topped up.
References by task
- Read Store PA broadcast workflow for slot cards, payloads, and review.
- Read Installation and authentication only when authorization needs attention.
- Read Installation registration for first-use registration.
- Read Tasks and results and Billing, errors, and recovery for task and billing facts.
- Read Bundled MCP Client diagnostics when the bundled client cannot connect.
- Read Automatic updates and safety for update controls.
- Read Uninstall and disconnect only when the user asks to remove the package.
Runtime and safe automatic updates
Use or invoke the bundled scripts/mcp_client.py for every Beatra operation.
Before ordinary commands it silently checks for a newer release at most once
every 24 hours per installation. Silent checks are enabled by default, and a
newer release installs without separate confirmation.
The updater accepts only the fixed official discovery address and immutable Beatra CDN path embedded for this package, channel, and locale. It verifies the discovery data, archive, manifest, and every file's size and checksum before replacement. It replaces only package-owned files and rejects redirects, downgrades, wrong package/channel/locale/version data, unexpected URLs, unsafe archives, and files outside the owned destination.
Update checks, downloads, verification, replacement, rollback, and recovery fail open: the current installation remains usable and the user's original command continues. An update failure never authorizes retrying a paid generation. The automatic-update choice persists across later commands for this installation:
python3 scripts/mcp_client.py update --auto off
python3 scripts/mcp_client.py update --auto on
python3 scripts/mcp_client.py update --check
Related skills
Fetch raw ad creative, app, ranking, and revenue data from AdMapix as structured JSON.
Stores durable facts in a categorized, plain-markdown vault on disk, alongside your agent's built-in memory.
Generate and edit Draw.io, Mermaid, and Excalidraw diagrams from natural language using a structured JSON spec.
Find why your productivity system keeps failing, then apply the smallest fix — capacity math, bottleneck routing, durable local notes.
Run Git operations — commits, branches, merges, rebases, conflict resolution, and recovery — with safety rules enforced.
More from beatra-ai
Browse all skillsTurn a topic description, a product photo, or brand references into a scroll-stopping event poster, promotional banner, or social media graphic. This AI poster maker and design generator creates photo-rich posters with strong visual hierarchy, category-matched styling, and text-safe zones for event, music, movie, product launch, and sale campaigns. Use it as an AI flyer maker for sale flyers, an event poster design tool, a movie poster generator, a music festival poster creator, a banner design AI for product launch posters, or a graphic design poster studio. Start from a campaign topic, transform a product or scene photo with style references, or refine an accepted draft toward a print-ready or share-ready result.
Turn a short video that already worked into your own version. Paste the link and this viral video teardown and short-video remake workflow reads the reference itself on TikTok, Douyin, Xiaohongshu, Instagram, YouTube, or X — caption, author, visible metrics, comments, and on YouTube the full transcript — or work from a file, screenshots, or your own description instead. It breaks the clip into its hook, body beats, and call to action, names the script pattern behind it, scores what carried the performance, then rewrites that structure around your product or topic — delivering a shot list with visuals and narration kept apart, reference frames from your own assets before any are generated, a narration track, and either an animated clip or segmented sources with a timecoded edit list. Use it to study a competitor's viral short, borrow a proven structure for Reels, Shorts, or WeChat Channels, rebuild a benchmark video under your own brand, or turn a saved reference into a content formula
Turn a comic panel, character sheet, webtoon frame, or frozen story beat into one dynamic comic-drama shot. This AI comic drama generator and motion comic maker animates an approved comic panel or manga frame into a short live shot, interpolates motion between a first and last comic panel, combines loose character, style, and scene references into a new comic-drama shot, or creates an original comic first frame and brings it to life. Use it for web-novel comic shots, motion comics, webtoon to video, manga panel animation, character entrances, emotional close-ups, action panels, dialogue reactions, and serialized creator workflows, with frozen character identity, approved art direction, and honest review of each independently generated shot.
Turn a novel chapter, web-novel excerpt, or story script into narrated vertical short video scenes with illustrated shots that keep every character looking the same from beat to beat. This AI story video maker pulls the hook and the beats out of your text, draws one shot for each, records the narration in a voice you choose, and sets each clip to its own narration, so a chapter arrives as an ordered set of scenes for story channels, book trailers, web-novel promotion, chapter recaps, and faceless storytelling accounts.
Turn written product answers and public note questions into one spoken wealth FAQ clip per question. This talking clip studio reads the questions buyers actually ask on public Xiaohongshu notes, records a speakable answer for each, and delivers separate short videos the advisor can post on their own.
Turn a user-supplied product factsheet and authorized stills into one wealth product talking clip per still. This product factsheet talking video studio writes a speakable product highlights talking clip for each photo, then animates a 2 to 15s product factsheet talking clip. Use it for wealth product talking pack and factsheet talking video pack work that stays one photo, one clip.