Gandr
Gandr is a cloud-based text-to-speech and voice cloning platform offering unlimited speech generation per concurrent stream, with no per-character usage fees. It supports 23 languages and provides instant voice cloning from approximately 10 seconds of reference audio, without per-voice training or additional charges. The service is designed for real-time applications: first audio byte in about 116 ms (p50) over the open internet, with streaming APIs over HTTP and WebSocket suitable for voice agents, dubbing, narration, and interactive systems. It integrates with tools like LiveKit, Pipecat, Vapi, Retell, Daily, and OpenAI's Whisper. Gandr emphasizes a flat pricing model based on concurrent streams, not usage volume, making costs predictable. It includes features like inaudible watermarking for provenance, and commits that user data is never used for training. The platform provides a console for testing voices and an exchange for browsing voice samples. The engine, Spex-TTS, is proprietary and offers three tiers of serving for reliability, with automatic failover and burst stream handling. Target use cases span audiobooks, e-learning, video dubbing, games, podcasts, and voice agents.
- Category
- AI Text-to-Speech
- Pricing
- Unknown
Problem solved
Traditional text-to-speech services charge per-character, leading to unpredictable costs for high-volume applications like voice agents or long-form content. Gandr solves this by offering unlimited speech per concurrent stream at a flat rate, enabling scalable, cost-effective audio generation without compromising on quality, speed, or language support.
Use cases
- Audiobook narration
- Video dubbing and localization
- E-learning content creation
- Interactive voice agents
- Game character voices
- Podcast production
Audience
- Developers
- Content Creators
- Enterprise
- Startups
Pros
- Unlimited usage per stream
- Flat pricing without per-character fees
- Fast real-time streaming (~116ms first audio)
- Instant voice cloning from short samples
- 23 languages with single voice cross-language capability
Cons
- Requires clean 10-second sample for cloning
- Pricing per stream may be high for single-user low-concurrency applications
- No offline or on-premise deployment
- Documentation on integration depth might be limited beyond key partners
Official website: https://gandr.ai/
Canonical LambdaBase page: https://www.lambdabase.com/tools/gandr