Announcer
Professional multilingual audio announcements, generated in seconds
The Problem
Every day, thousands of venues struggle with getting the right message to the right audience, in the right language, at the right time. A karate tournament needs to call competitors to the mat with names pronounced correctly. A bus station needs to announce a platform change in four languages. The traditional process is painful — record with voice actors, edit in audio software, translate, re-record, upload. What should take seconds takes days. What should cost nothing costs thousands.
The Solution
Announcer is a self-hosted platform that turns text into broadcast-ready audio across 74 languages. Compose announcements visually from reusable templates, preview them instantly in the browser, and deploy to any speaker system — web player, local PA, or REST API. No recording studios, no voice actors, no audio engineers required.
Key Capabilities
Compose Like Building Blocks
Announcements are assembled from reusable segments — speech blocks and sound effects — arranged in sequence. Templates use Liquid syntax for dynamic content, populated from live data sources connected to your existing systems.
Generate in Every Language at Once
Toggle "Speak All Languages" on a channel and every announcement is automatically generated in all configured languages, concatenated into a single audio file with natural pauses. Write once, broadcast everywhere.
Multiple TTS Providers
Choose ElevenLabs (74 languages, emotion tags), OpenAI gpt-4o-mini-tts, self-hosted Chatterbox for air-gapped environments, or macOS Say for local development. Each channel can use a different provider independently.
Smart Caching & Deduplication
Chunk-level TTS responses are cached for one week. An MD5 hash of the full segment composition prevents duplicate generation — identical announcements are cloned instantly with no API call. Reduces costs by 40–60% in high-repetition environments.
Multi-Tenancy
A single Announcer instance serves multiple independent tenants, each with their own channels, templates, and users. Duplicate an entire tenant — template library, data sources, sound files — with one click to spin up a new venue.
Three Player Types
Web Player (React-based browser player), Local Player (server-side audio output via PortAudio for PA systems), and External Player (REST API for kiosks, digital signage, and custom devices).
How It Works
Click templates from the sidebar to add segments to the queue
For templates with data sources, pick from a searchable list of live options
Hear exactly what your audience will hear before committing
Audio generation begins — segments are saved and the queue clears
The announcement is delivered to all configured player endpoints
Traditional Recording vs. Announcer
One platform. 74 languages. Seconds, not days.
