Multilingual AI Mobile Platform
FarsiTalk, FieldTalk and InstaSpeak
One Flutter platform supporting three multilingual voice products: FarsiTalk, FieldTalk and InstaSpeak. It combines live speech recognition, translation, neural voice generation, entitlements, subscriptions, privacy controls and mobile release operations.
Speak in one language, hear another.
- 1. Spoken input
- 2. Speech recognition and translation
- 3. Cloud orchestration
- 4. Translated voice response
- Role
- Product, architecture and mobile delivery
- Serves
- People who need real-time spoken translation across languages and contexts.
- Core technologies
- Flutter · Cloud Run · Firebase · Azure AI
Executive overview
The problem. Building separate voice-translation apps for different markets would multiply cost and maintenance. The goal was a shared platform that still lets each product keep its own identity and customer experience.
Who it serves. People who need real-time spoken translation across languages and contexts.
What Daynis did. Product, architecture and mobile delivery. One Flutter platform supporting three multilingual voice products: FarsiTalk, FieldTalk and InstaSpeak. It combines live speech recognition, translation, neural voice generation, entitlements, subscriptions, privacy controls and mobile release operations.
Why it matters. Reuses a robust platform across multiple products and markets while retaining distinct customer experiences and product identities.
Business architecture and impact
Capabilities created
- Live speech recognition, translation and neural voice output
- Subscriptions and entitlements with server-enforced quotas
- Three distinct products from one platform
- Automated iOS and Android releases
Reuses a robust platform across multiple products and markets while retaining distinct customer experiences and product identities.
InstaSpeak
Voice translation product on the same platform.
System architecture
- 01
FarsiTalk, FieldTalk, InstaSpeak users
Where requests and source data originate.
- 02
Flutter apps (shared core)
The interface people use day to day.
- 03
Cloud Run streaming services
Orchestrates requests and enforces business rules.
- 04
Speech recognition + Translation
Models applied under defined, reviewable constraints.
- 05
Firebase
The governed system of record.
Key flows
- FarsiTalk, FieldTalk, InstaSpeak users to Flutter apps (shared core)
- Flutter apps (shared core) to Cloud Run streaming services (audio stream)
- Cloud Run streaming services to Speech recognition
- Speech recognition to Translation
- Translation to Neural text-to-speech
Technical depth
Highlights
- Flutter with Riverpod and GoRouter
- Real-time streaming through Cloud Run and Firebase
- Google and Azure AI services for speech, translation and neural TTS
- Authentication and App Check
- iOS and Android release automation
Architecture decisions
One platform, three products
A shared core carries streaming, AI integration and entitlements; each product configures its own identity and experience.
Multi-cloud AI services
Google and Azure services are combined so each capability uses the provider best suited to it.
Quotas on the server
Usage limits are enforced server-side rather than trusted to the device.
Security and governance
- Authentication and App Check protect backend services
- Server-enforced quotas and entitlements
- Privacy controls built into the product
What I led
- Product and commercial decision support across three products
- Architecture ownership of the shared platform
- Mobile release engineering standards
Evidence and links
Evidence: Public productPublicly available product.
Related speaking topic
Building something similar?
Talk to Daynis about ai and mobile products architecture and delivery.
Start a conversation