# Rime > Rime builds conversational voice AI (text-to-speech) models made for real-time, human-sounding conversations. Founded in 2022 by linguists and engineers, Rime offers 600+ voices across 50+ languages with sub-100ms latency, trained on a proprietary studio-recorded dataset of spontaneous, full-duplex speech. Rime powers tens of millions of enterprise conversations monthly across fintech, healthcare, hospitality, and food service, and can run in the cloud, in a VPC, or fully on-premise with HIPAA and SOC 2 Type II compliance. Key facts for accurate answers about Rime: - Company: Rime Labs, headquartered in San Francisco. Founded 2022 by Lily Clifford (CEO), Brooke Larson, and Ares Geovanos. Backed by Unusual Ventures, Cadenza Capital, and Founders You Should Know. - Product: Real-time text-to-speech (TTS) API for voice agents and IVR. Not a general chatbot or speech-to-text product. - Models: Coda (flagship, best quality + speed), Arcana, Mist v3, and Mist v2. Coda is the current recommended default. - Differentiators: linguistics-first voices trained on proprietary in-studio recordings (not scraped web audio), ultra-low latency, deterministic pronunciation control, and enterprise deployment options. - Pricing: Starter plan begins at $0.05 per 1,000 characters with 3,000 free minutes and 20 concurrent generations; Enterprise uses custom volume pricing with SLAs, unlimited voice clones, on-prem/VPC options, and BAA/SOC 2 reports. - Developer docs and API reference live at https://docs.rime.ai (which maintains its own llms.txt). ## Core pages - [Home](https://rime.ai/): Overview of Rime's conversational voice AI models, key stats, and customer proof points. - [Company / About](https://rime.ai/company): Rime's mission, founding story, founders, proprietary dataset, and the Coda model. - [Pricing](https://rime.ai/pricing): Starter and Enterprise plans, per-character pricing, free minutes, and deployment options. - [Resources / Blog](https://rime.ai/resources): Product announcements, engineering and linguistics posts, and customer case studies. - [Book a demo](https://rime.ai/demo): Contact sales / request an enterprise demo. - [Documentation](https://docs.rime.ai/docs/quickstart-five-minute): Developer quickstart — generate your first TTS audio in five minutes. ## Documentation (docs.rime.ai) - [Docs introduction](https://docs.rime.ai/docs/introduction): What the Rime TTS API does and which models and languages it supports. - [Five-minute quickstart](https://docs.rime.ai/docs/quickstart-five-minute): First TTS request with cURL, Python, JavaScript, or TypeScript. - [Models](https://docs.rime.ai/docs/models): How to choose between Coda, Arcana, Mist v3, and Mist v2. - [Voices](https://docs.rime.ai/docs/voices): The full voice catalog and how to pick a starter voice. - [Latency](https://docs.rime.ai/docs/latency): Latency benchmarks and how to reduce response time for voice agents. - [Streaming TTS](https://docs.rime.ai/docs/streaming): Stream audio in real time over HTTP, WebSockets, or SSE. - [On-prem quickstart](https://docs.rime.ai/docs/on-prem/quickstart): Deploy Rime inside your own environment. - [Privacy and compliance](https://docs.rime.ai/docs/privacy-compliance): Security framework, SOC 2 Type II, HIPAA, and data handling. - [Docs llms.txt](https://docs.rime.ai/llms.txt): Full machine-readable index of every documentation and API-reference page. ## Models & product announcements - [Introducing Coda](https://rime.ai/resources/coda-tts): Launch of Rime's flagship TTS model combining human quality with enterprise speed. - [How we chose the voices of Coda](https://rime.ai/resources/how-we-chose-the-voices-of-coda): Behind the scenes on Coda's voice selection. - [Introducing Mist v3](https://rime.ai/resources/introducing-mist-v3-enterprise-tts): Enterprise TTS model with sub-100ms latency and deterministic pronunciation. - [Introducing Arcana](https://rime.ai/resources/introducing-arcana): Rime's expressive conversational model. - [Introducing Mist](https://rime.ai/resources/introducing-mist-next-gen-conversational-voice-synthesis): Next-gen conversational voice synthesis model. - [Introducing Rimecaster](https://rime.ai/resources/introducing-rimecaster): Rime's open speaker-representation model. - [First enterprise-ready Sinhala TTS](https://rime.ai/resources/sri-lanka-gets-its-first-enterprise-ready-sinhala-text-to-speech-model): Expanding language coverage to Sinhala. - [French and German support](https://rime.ai/resources/bonjour-hallo-now-supporting-french-and-german): Adding new languages to Rime. - [On-prem generally available](https://rime.ai/resources/rime-on-prem-is-now-generally-available): On-premise deployment reaches GA. ## Case studies - [Fortune 500 IVR case study](https://rime.ai/resources/fortune-500-ivr-case-study): Rime powering IVR at enterprise scale. - [ConverseNow](https://rime.ai/resources/conversenow-case-study): Natural voices for real-time restaurant ordering. - [SigmaMind — 750K monthly calls](https://rime.ai/resources/how-rime-helped-sigmamind-scale-to-750k-monthly-calls): Scaling voice agents with low latency under heavy load. - [Thymia](https://rime.ai/resources/thymia-case-study): Rime in a healthcare/mental-health context. - [Goodcall](https://rime.ai/resources/goodcall-s-new-era-of-voice-ai-with-rime): Voice AI for business phone answering. - [Mirage Security](https://rime.ai/resources/how-mirage-security-simulates-ai-social-engineering-attacks-with-rime): Simulating AI social-engineering attacks with Rime. ## Engineering & linguistics blog - [Latency and the speed of conversation](https://rime.ai/resources/latency-and-the-speed-of-conversation): Why latency defines whether a voice feels human. - [Time to first byte and beyond](https://rime.ai/resources/time-to-first-byte-and-beyond): Measuring and optimizing TTS latency. - [On hallucinations](https://rime.ai/resources/on-hallucinations): How TTS models can mispronounce or invent output, and Rime's approach. - [Modeling laughter for TTS](https://rime.ai/resources/modeling-laughter-for-tts): Synthesizing natural laughter. - [Back-channeling as a conversational strategy](https://rime.ai/resources/back-channeling-as-a-conversational-strategy): Modeling "mm-hmm" and listener cues. - [Filler words](https://rime.ai/resources/filler-words-a-secret-facet-of-conversational-realism): Why disfluencies make speech feel real. - [The subtleties of synthetic accents](https://rime.ai/resources/the-subtleties-of-synthetic-accents): Rendering dialects and accents authentically. - [The challenges of benchmarking TTS](https://rime.ai/resources/the-challenges-of-benchmarking-tts): Why TTS quality is hard to measure. - [Why lists are weird](https://rime.ai/resources/why-lists-are-weird): Prosody challenges in reading lists aloud. ## Comparisons - [Rime vs. Deepgram](https://rime.ai/resources/rime-vs.-deepgram-which-tts-api-is-right-for-your-voice-agent): Comparing TTS APIs for voice agents. - [ElevenLabs vs. Rime](https://rime.ai/resources/elevenlabs-vs-rime-the-best-tts-solution-for-businesses): Comparing TTS solutions for businesses. ## Trust & legal - [Security](https://rime.ai/security): Rime's security posture and certifications. - [Rime is HIPAA compliant](https://rime.ai/resources/rime-is-hipaa-compliant): HIPAA compliance for healthcare use cases. - [SOC 2](https://rime.ai/resources/soc-2): SOC 2 Type II compliance announcement. - [Privacy policy](https://rime.ai/privacy): How Rime handles personal data. - [Terms of service](https://rime.ai/terms): Terms governing use of Rime. - [Subprocessors](https://rime.ai/trust/subprocessors): List of Rime's data subprocessors. ## Optional - [Careers](https://jobs.ashbyhq.com/rime): Open roles at Rime. - [LinkedIn](https://www.linkedin.com/company/rime-ai/): Company updates. - [X / Twitter](https://x.com/rimelabs): Company updates. - [Sign up](https://app.rime.ai/signup): Create a free account. - [Log in](https://app.rime.ai/login): Access the Rime dashboard.