Cartesia vs Speechmatics: pricing, features and which to choose

Cartesia is a San Francisco startup founded in 2023 by Karan Goel, Albert Gu, Arjun Desai, Brandon Yang and Chris Ré, researchers from Stanford's AI lab. Gu co-created the S4 and Mamba state space model architectures, which process long sequences with less memory than transformers, and that research is the basis of the company's products. Cartesia launched its Sonic text-to-speech model in 2024 and raised a $64 million Series A led by Kleiner Perkins in March 2025, bringing total funding to $91 million, according to Fortune. Its homepage names ServiceNow, Retell, Blue Machines and Quora among customers. The company's pitch is speed: models that start speaking quickly enough for live phone conversation, sold through one API that covers speech generation, transcription and complete voice agents.

Developers reach Cartesia through a REST and WebSocket API or the browser playground. Sonic turns text into streaming audio, with controls for speed, emotion and pronunciation, and supports more than 40 languages. Instant voice cloning creates a voice from a few seconds of audio on paid plans, and professional cloning trains a closer copy from longer recordings on the Startup plan and above. A localisation feature gives a cloned voice a new accent. Ink transcribes live audio for agents that need to hear callers. Managed Agents, the newest product, hosts complete phone agents with telephony, so a team can define an agent's prompt and tools without wiring a separate orchestration layer. Cartesia also offers on-premises and on-device deployment for customers with data residency rules.

Cartesia's voice library and creator tools are thinner than ElevenLabs, which also offers dubbing, sound effects and music, so it is a weaker fit for video producers and audiobook narrators. Managed Agents are newer than Vapi or Retell and have fewer integrations. Credit-based billing means long-form narration gets expensive faster on the low tiers. ElevenLabs Flash competes directly on latency with a larger voice catalog. Deepgram's Aura voices are cheaper per minute but less expressive, and Hume, once a rival on expressive speech, is closing its TTS API in November 2026. Pick Cartesia when you are building a real-time voice agent or interactive app and the delay before the first syllable is the metric you care about, or when you need on-device or on-premises speech models.

Speechmatics grew out of research by Tony Robinson, who applied neural networks to speech recognition at Cambridge University in the 1980s, and the company was founded in Cambridge in 2006. Katy Wigdahl is chief executive. In 2022 it raised a $62 million Series B led by Susquehanna Growth Equity, according to TechCrunch. Its stated aim is to understand every voice, and its pitch rests on accuracy for accents, dialects and non-native speakers that trip up other engines. Customers range from broadcasters running live captions to contact centres, legal transcription firms and medical documentation companies. Speechmatics is one of the few speech vendors that has always sold on-premises and on-device deployment alongside its cloud API, which keeps it on shortlists for banks, governments and hospitals with strict data rules.

The product line now covers three jobs. Batch transcription processes recorded audio and video files, with speaker diarization, punctuation, custom dictionaries and translation. Real-time transcription streams live audio for captions and analytics. Agent STT, launched on September 17, 2026, is built for voice agents: it returns transcripts with turn detection, live speaker labels, speaker identification and up to 1,000 custom vocabulary terms over one connection. It plugs into the LiveKit and Pipecat agent frameworks, and Speechmatics says early users integrated it in under an hour. A text-to-speech API, added in June 2026, gives voice agents a matching output voice. Developers use REST and WebSocket APIs with SDKs, or a web portal for one-off files, and add-ons such as summaries and sentiment are priced separately.

Speechmatics' pricing is harder to read than Deepgram's or AssemblyAI's, because the published rates bundle discounts that depend on data sharing and annual commitments. Its text-to-speech is new and has a smaller voice range than ElevenLabs or Cartesia. Deepgram is the closest competitor for voice agents, with Flux turn detection at similar prices and a larger developer community. AssemblyAI is strong on audio intelligence features for recorded calls and meetings. OpenAI's transcription models are convenient for teams already on its API but offer no on-premises option. Pick Speechmatics for accented or code-switched speech, for medical transcription, for live broadcast captioning in many languages, or when audio must stay inside your own data centre. Benchmark it against Deepgram on your own recordings before signing a volume deal.