News
Explore five alternatives to ElevenLabs for AI phone agents, comparing costs, concurrency, and telephony compatibility.
As companies seek alternatives to ElevenLabs for AI-powered phone agents, understanding the technical and operational changes involved is critical. This guide evaluates five text-to-speech (TTS) providers—AssemblyAI, Inworld, Cartesia, Deepgram, and Rime—focusing on their pricing, concurrency models, and telephony compatibility.
Why Companies Are Considering Alternatives
Several factors have prompted organizations to explore replacements for ElevenLabs. The platform’s session-based concurrency limits, idle connection timeouts, and its removal from LiveKit Inference on August 31, 2026, are notable constraints. For example, ElevenLabs’ Text to Dialogue WebSocket connections are metered as dialogue sessions, with limits based on plan tiers. Open connections past the limit are rejected, and idle connections close after 20 seconds unless kept alive. These restrictions can hinder scalability for phone agents handling high call volumes or long pauses.
Key Alternatives and Their Features
AssemblyAI
AssemblyAI-TTS-1 offers a concurrency model based on call starts rather than open calls, which can simplify scaling. It supports native pcm_mulaw and pcm_alaw at 8 kHz and provides US and EU data zones. Pricing is $15 per 1 million characters with no commitment. However, it supports only six languages (English, Spanish, French, German, Italian, and Portuguese) and lacks advanced controls like speed, pitch, and emotion parameters. In benchmark tests, AssemblyAI-TTS-1 achieved 93.7% entity accuracy, comparable to ElevenLabs v4 at 93.6%.
Inworld
Inworld’s TTS-2 and TTS-2 Flash models cover over 200 languages and locales. Concurrency is measured in generations, with limits ranging from 5 on the On-Demand plan to 500 on the Growth plan. WebSocket connections are capped separately at 10 times the generation limit. Pricing starts at $25 per 1 million characters for On-Demand users.
Cartesia
Cartesia’s Sonic 3.6 supports 44 languages and outputs pcm_mulaw for telephony. Pricing is credit-based, with plans ranging from $5 per month for 100,000 credits to $299 per month for 8 million credits. Concurrency limits vary by plan, from 2 requests on the Free tier to 15 on the Scale tier, with custom limits available for enterprise users.
Deepgram
Deepgram’s Flux TTS and Aura-2 models stream over WebSocket and support mulaw and alaw at 8 kHz. Concurrency is fixed, with up to 45 simultaneous streaming requests on the Pay-as-You-Go plan. Aura-2, introduced in April 2025, boasts sub-200ms time to first byte and costs $0.030 per 1,000 characters.
Rime
Rime offers Mist v3 and Coda v2 with pricing based on estimated minutes rather than characters. Mist v3 costs $0.030 per minute, while Coda v2 is priced at $0.050. Concurrency limits depend on the plan, with Starter tiers supporting five concurrent generations.
Considerations for Migration
Switching TTS providers often requires updates to configuration files, API credentials, and output parameters. For example, LiveKit and Pipecat users may need to replace plugins or services and adjust settings for sample rates and interruption behavior. Raw Twilio implementations demand additional steps, such as handling streaming protocols and audio conversion.
Companies evaluating alternatives should calculate monthly costs based on actual call transcripts, test end-to-end latency, and ensure compatibility with telephony requirements like audio/x-mulaw at 8,000 Hz.
Credit: Source link
