Text to Speech APIs
Text-to-speech providers differ in voice range, latency and extras. ElevenLabs offers thousands of voices across many languages plus cloning and dubbing, Cartesia focuses on low latency over HTTP, SSE or WebSocket, and Fish Audio supports emotion tags in the text. IBM Watson Text to Speech and Audexum also run freemium plans with voices in many languages. Free AI APIs →
| API | Pricing | Auth | Format | Free tier | Status |
|---|---|---|---|---|---|
AudexumText-to-speech in 43 voices and 32 languages, plus speech-to-text transcription in 100+ languages. AI & Machine Learning › Speech | Freemium | API key | JSON | — | Live |
CartesiaLow-latency text-to-speech over HTTP, SSE or WebSocket, plus speech-to-text and instant voice cloning. AI & Machine Learning › Speech | Freemium | API key | Binary | — | Live |
DeepgramSpeech-to-text and text-to-speech transcription with streaming and pre-recorded audio, plus voice agents. AI & Machine Learning › Speech | Free trial | API key | JSON | — | Live |
ElevenLabsText-to-speech in 70+ languages with 10,000+ voices, plus speech-to-text, voice cloning and dubbing. AI & Machine Learning › Speech | Freemium | API key | JSON | — | Live |
Fish AudioText-to-speech with emotion tags, instant voice cloning from short clips, and transcription, billed per byte. AI & Machine Learning › Speech | Freemium | API key | Binary | — | Live |
IBM Watson Text to SpeechConverts text into natural-sounding speech in many languages and voices via IBM Watson. AI & Machine Learning › Speech | Freemium | API key | Binary | — | Untested |
AudexumFreemiumLive
Text-to-speech in 43 voices and 32 languages, plus speech-to-text transcription in 100+ languages.
API key · JSON
—
llms.txt
CartesiaFreemiumLive
Low-latency text-to-speech over HTTP, SSE or WebSocket, plus speech-to-text and instant voice cloning.
API key · Binary
—
no machine-readable spec
DeepgramFree trialLive
Speech-to-text and text-to-speech transcription with streaming and pre-recorded audio, plus voice agents.
API key · JSON
—
no machine-readable spec
ElevenLabsFreemiumLive
Text-to-speech in 70+ languages with 10,000+ voices, plus speech-to-text, voice cloning and dubbing.
API key · JSON
—
no machine-readable spec
Fish AudioFreemiumLive
Text-to-speech with emotion tags, instant voice cloning from short clips, and transcription, billed per byte.
API key · Binary
—
OpenAPI spec
IBM Watson Text to SpeechFreemiumUntested
Converts text into natural-sounding speech in many languages and voices via IBM Watson.
API key · Binary
—
no machine-readable spec
Nothing matches that combination. Clear the filters and start from the category.
Choosing a Text to Speech API
Is there a free Text to Speech API?
Yes. ElevenLabs, Cartesia, IBM Watson Text to Speech, Audexum and Fish Audio have a free tier alongside paid plans. See all free AI APIs.
Which Text to Speech APIs work without an API key?
All of them need a key or account. ElevenLabs, Cartesia, IBM Watson Text to Speech, Audexum and Fish Audio give one out free when you sign up.
Can I use Text to Speech APIs in a commercial project?
Deepgram permits commercial use. ElevenLabs, Cartesia, Audexum and Fish Audio restrict it, for example to non-commercial or attributed use, so read the terms first.
Which Text to Speech APIs work with AI agents?
Fish Audio publishes an OpenAPI spec for generating clients and tools.
Broader: Speech APIs · AI & Machine Learning APIs
More in Speech: Speech to Text APIs · Voice Cloning APIs