AI Voice & Speech Tools

Voice is becoming the primary interface for software, and the tooling has finally caught up. This collection gathers the APIs, models, and products that turn text into lifelike speech, clone a voice from a short sample, transcribe conversations in real time, and power voice-first agents. Whether you're building a phone-based AI receptionist, adding narration to a product, or shipping a dictation workflow, these are the building blocks worth knowing.

We prioritized tools with real developer surfaces — documented APIs, SDKs, or MCP support — over demo-only voice toys. ElevenLabs and Deepgram anchor the list as the two most widely adopted synthesis and transcription APIs respectively, while Vapi, Retell, and Cartesia cover the fast-growing voice-agent stack. A few picks (Krisp, VoiceDrop, Willow) show how voice is being applied to specific jobs like accent conversion, outbound sales, and team dictation. We left out general-purpose AI platforms that happen to include a TTS endpoint, since the focus here is voice as a product category.