# Gradium

Research, engineering, and product updates on voice AI from Gradium.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Gradium MCP beta

DevFeed: [Gradium MCP beta](<https://devfeed.tech/articles/gradium-mcp-beta-144139.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/gradium-mcp-beta>)

Author: Gradium

Published: 2026-11-19T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Claude Code](<https://devfeed.tech/topics/claude-code.md>), [MSP MCP](<https://devfeed.tech/topics/msp-mcp.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [account](<https://devfeed.tech/tags/account.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [ai-skills](<https://devfeed.tech/tags/ai-skills.md>), [ai-voice-generation](<https://devfeed.tech/tags/ai-voice-generation.md>), [api-key](<https://devfeed.tech/tags/api-key.md>), [audio](<https://devfeed.tech/tags/audio.md>), [audio-files](<https://devfeed.tech/tags/audio-files.md>), [beta](<https://devfeed.tech/tags/beta.md>), [claude-code-mcp](<https://devfeed.tech/tags/claude-code-mcp.md>), [claude-code-plugin](<https://devfeed.tech/tags/claude-code-plugin.md>), [codex-mcp](<https://devfeed.tech/tags/codex-mcp.md>), [codex-plugin](<https://devfeed.tech/tags/codex-plugin.md>), [coding-agent](<https://devfeed.tech/tags/coding-agent.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [gradium-mcp](<https://devfeed.tech/tags/gradium-mcp.md>), [mcp](<https://devfeed.tech/tags/mcp.md>), [mcp-server](<https://devfeed.tech/tags/mcp-server.md>), [model-context-protocol](<https://devfeed.tech/tags/model-context-protocol.md>), [one](<https://devfeed.tech/tags/one.md>), [product](<https://devfeed.tech/tags/product.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [script](<https://devfeed.tech/tags/script.md>), [session](<https://devfeed.tech/tags/session.md>), [sign-in](<https://devfeed.tech/tags/sign-in.md>), [speech](<https://devfeed.tech/tags/speech.md>), [text-to-speech-mcp](<https://devfeed.tech/tags/text-to-speech-mcp.md>), [tts-mcp-server](<https://devfeed.tech/tags/tts-mcp-server.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-selection](<https://devfeed.tech/tags/voice-selection.md>)

### AI overview

Gradium MCP is a beta that lets Claude Code or Codex browse Gradium's voice catalog and generate speech within a coding session. Setup uses the MCP endpoint and requires a Gradium account sign-in, but no API key; if tools do not appear after sign-in, the article advises starting a new session and repeating setup. Voice browsing is free, while audio generation uses credits at the same rate as the API. The beta covers text-to-speech and voice selection; speech-to-text and voice design are only possible next candidates, with priorities to be guided by feedback.

### Source excerpt

The Gradium MCP is available today in beta. Your coding agent can choose a Gradium voice and turn a script into speech in the same session where it writes your code. You can sign in with your Gradium account, so you don't need an API key. Paste one prompt to set it up.

## A Closer Look at Phonon

DevFeed: [A Closer Look at Phonon](<https://devfeed.tech/articles/a-closer-look-at-phonon-121146.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/phonon-closer-look>)

Author: Gradium; Hugging Face

Published: 2026-11-05T00:00:00Z

Content type: article

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Open Source](<https://devfeed.tech/topics/open-source.md>), [hugging face](<https://devfeed.tech/topics/hugging-face.md>), [spaces](<https://devfeed.tech/topics/spaces.md>), [huggingface\_hub](<https://devfeed.tech/topics/huggingface-hub.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [kernels](<https://devfeed.tech/topics/kernels.md>), [open-llm-leaderboard](<https://devfeed.tech/topics/open-llm-leaderboard.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [apache-2-0](<https://devfeed.tech/tags/apache-2-0.md>), [apache-2-0-tts](<https://devfeed.tech/tags/apache-2-0-tts.md>), [apple](<https://devfeed.tech/tags/apple.md>), [apple-watch](<https://devfeed.tech/tags/apple-watch.md>), [apps](<https://devfeed.tech/tags/apps.md>), [arabic](<https://devfeed.tech/tags/arabic.md>), [arm](<https://devfeed.tech/tags/arm.md>), [audio](<https://devfeed.tech/tags/audio.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [backends](<https://devfeed.tech/tags/backends.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [browser](<https://devfeed.tech/tags/browser.md>), [chat-template](<https://devfeed.tech/tags/chat-template.md>), [checkpoint](<https://devfeed.tech/tags/checkpoint.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [codec](<https://devfeed.tech/tags/codec.md>), [hugging-face](<https://devfeed.tech/tags/hugging-face.md>), [multilingual-tts](<https://devfeed.tech/tags/multilingual-tts.md>), [on-device-tts](<https://devfeed.tech/tags/on-device-tts.md>), [open-tts-leaderboard](<https://devfeed.tech/tags/open-tts-leaderboard.md>), [open-weight-tts](<https://devfeed.tech/tags/open-weight-tts.md>), [phonon](<https://devfeed.tech/tags/phonon.md>), [phonon-31m](<https://devfeed.tech/tags/phonon-31m.md>), [phonon-40m](<https://devfeed.tech/tags/phonon-40m.md>), [pocket-tts](<https://devfeed.tech/tags/pocket-tts.md>), [ptts](<https://devfeed.tech/tags/ptts.md>), [time-to-first-audio](<https://devfeed.tech/tags/time-to-first-audio.md>), [transformers-tts](<https://devfeed.tech/tags/transformers-tts.md>), [tts](<https://devfeed.tech/tags/tts.md>)

### AI overview

Phonon is an open-weights text-to-speech model designed to run locally without a server or network, keeping typed text on the device. The release includes models of about 65 MB and 75 MB, more than 100 built-in voices, and support across English, French, German, Spanish, and Portuguese. Phonon 31m is Apache 2.0, including commercial use; Phonon 40m is non-commercial. The authors attribute the models' compact quality to flow matching and report that Phonon ranks first in German and Spanish intelligibility and in the top three in each evaluated language, while leading CPU streaming time-to-first-audio benchmarks. Their ptts framework targets single-stream inference across CPU, GPU, and NPU backends; the Transformers implementation targets GPU throughput with batched generation and full-graph torch.compile. Custom voices can be cloned from a short recording with the speaker's consent or designed in text. The leaderboard is described as new, and the reported rankings apply to the languages evaluated.

### Source excerpt

Two weeks after opening up Phonon, the story of the launch: why we built it, what it can do, the demos that came with it, and how it did on Hugging Face's Open TTS Leaderboard.

## Gradium and Yampa partner on voice AI agents for customer experience

DevFeed: [Gradium and Yampa partner on voice AI agents for customer experience](<https://devfeed.tech/articles/gradium-and-yampa-partner-on-voice-ai-agents-for-customer-experience-101270.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/yampa-gradium-partnership>)

Author: Gradium

Published: 2026-10-26T00:00:00Z

Content type: article

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [AI for customer service automation](<https://devfeed.tech/topics/ai-for-customer-service-automation.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [The Three Main Cloud Service Models](<https://devfeed.tech/topics/the-three-main-cloud-service-models.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-platform](<https://devfeed.tech/tags/agentic-platform.md>), [agentic-platform-voice](<https://devfeed.tech/tags/agentic-platform-voice.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [build](<https://devfeed.tech/tags/build.md>), [customer](<https://devfeed.tech/tags/customer.md>), [customer-experience](<https://devfeed.tech/tags/customer-experience.md>), [customer-experience-voice-ai](<https://devfeed.tech/tags/customer-experience-voice-ai.md>), [deploy](<https://devfeed.tech/tags/deploy.md>), [enterprises](<https://devfeed.tech/tags/enterprises.md>), [eu](<https://devfeed.tech/tags/eu.md>), [eu-data-residency-voice-ai](<https://devfeed.tech/tags/eu-data-residency-voice-ai.md>), [experience](<https://devfeed.tech/tags/experience.md>), [french-voice-agent](<https://devfeed.tech/tags/french-voice-agent.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [on-premise](<https://devfeed.tech/tags/on-premise.md>), [on-premise-voice-agents](<https://devfeed.tech/tags/on-premise-voice-agents.md>), [partner](<https://devfeed.tech/tags/partner.md>), [platform](<https://devfeed.tech/tags/platform.md>), [product](<https://devfeed.tech/tags/product.md>), [run](<https://devfeed.tech/tags/run.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-agents-customer-care](<https://devfeed.tech/tags/voice-agents-customer-care.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>), [voice-cloning-brand-voice](<https://devfeed.tech/tags/voice-cloning-brand-voice.md>), [voice-designer](<https://devfeed.tech/tags/voice-designer.md>), [yampa-gradium](<https://devfeed.tech/tags/yampa-gradium.md>), [yampa-partnership](<https://devfeed.tech/tags/yampa-partnership.md>), [yampa-voice-engine](<https://devfeed.tech/tags/yampa-voice-engine.md>)

### AI overview

Gradium's Text-to-Speech and Speech-to-Text run inside Yampa's Voice Engine, which orchestrates calls including turn-taking, interruptions, silences and background noise. Organizations can deploy the combined platform as SaaS on European infrastructure or on-premise; the source says conversations stay within an organization-controlled perimeter and lists Gradium's ISO 27001 and SOC 2 Type II certifications and Yampa's SOC 2 and GDPR compliance. Gradium highlights sub-50 ms speech onset, pronunciation controls for structured data and specialist terms, and French voices including regional accents. Yampa provides tools to design, test, monitor and govern agents across voice and other channels, with CRM, helpdesk and telephony integrations. The companies say they use customer-call feedback to improve models within weeks. Yampa reports clients resolve 65% of after-hours calls autonomously and cut escalations by 40% in two months; these are company-reported results.

### Source excerpt

Gradium's voice models now run natively inside Yampa, the agentic platform enterprises use to build, deploy and supervise customer care agents, in SaaS on EU infrastructure or on-premise.

## Gradium releases Phonon, an open-source text-to-speech model that runs on any device

DevFeed: [Gradium releases Phonon, an open-source text-to-speech model that runs on any device](<https://devfeed.tech/articles/gradium-releases-phonon-an-open-source-text-to-speech-model-that-runs-on-any-device-92452.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/gradium-releases-phonon>)

Author: Gradium

Published: 2026-10-21T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Open Source](<https://devfeed.tech/topics/open-source.md>), [llama.cpp](<https://devfeed.tech/topics/llama-cpp.md>)

Tags: [accessibility](<https://devfeed.tech/tags/accessibility.md>), [accessibility-tools](<https://devfeed.tech/tags/accessibility-tools.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [amd](<https://devfeed.tech/tags/amd.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [apache-2-0-tts](<https://devfeed.tech/tags/apache-2-0-tts.md>), [browser](<https://devfeed.tech/tags/browser.md>), [device](<https://devfeed.tech/tags/device.md>), [embedded](<https://devfeed.tech/tags/embedded.md>), [gpus](<https://devfeed.tech/tags/gpus.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [gradium-phonon](<https://devfeed.tech/tags/gradium-phonon.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [laptops](<https://devfeed.tech/tags/laptops.md>), [model](<https://devfeed.tech/tags/model.md>), [multilingual-tts](<https://devfeed.tech/tags/multilingual-tts.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-cuda-tts](<https://devfeed.tech/tags/nvidia-cuda-tts.md>), [nvidia-gpus](<https://devfeed.tech/tags/nvidia-gpus.md>), [nvidia-gtc](<https://devfeed.tech/tags/nvidia-gtc.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [on-device-tts](<https://devfeed.tech/tags/on-device-tts.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [open-source-text-to-speech](<https://devfeed.tech/tags/open-source-text-to-speech.md>), [phones](<https://devfeed.tech/tags/phones.md>), [phonon](<https://devfeed.tech/tags/phonon.md>), [speech](<https://devfeed.tech/tags/speech.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [voice-cloning](<https://devfeed.tech/tags/voice-cloning.md>), [voice-design](<https://devfeed.tech/tags/voice-design.md>), [webgpu-tts](<https://devfeed.tech/tags/webgpu-tts.md>)

### AI overview

Gradium's Phonon is an open-source text-to-speech model designed to generate speech locally, without a network connection or cloud dependency. The release describes 31M- and 40M-parameter variants, supporting English, French, German, Spanish, and Portuguese, with CPU, CUDA, and WebGPU inference paths and integrations for Rust, Python, Swift, and browsers. Gradium reports that blind human evaluations favored Phonon over Kokoro, Magpie, and Pocket TTS, and that on an NVIDIA L40S its CUDA backend produced first audio in 5 ms versus 43 ms on an AMD EPYC 7413 CPU. Weights and inference code are released under Apache 2.0; voice cloning and voice design are offered through Gradium's API. These performance and evaluation results are claims in the supplied release text.

### Source excerpt

Trained on NVIDIA GPUs, Phonon generates natural speech locally in as little as 64.5 MB, on phones, laptops, embedded hardware and in the browser.

## Open-Sourcing Phonon: A Suite of Frontier On-Device TTS Models

DevFeed: [Open-Sourcing Phonon: A Suite of Frontier On-Device TTS Models](<https://devfeed.tech/articles/open-sourcing-phonon-a-suite-of-frontier-on-device-tts-models-91160.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/release-phonon-open-weight>)

Author: Gradium

Published: 2026-10-20T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Ollama](<https://devfeed.tech/topics/ollama.md>), [Transformers](<https://devfeed.tech/topics/transformers.md>), [gemma](<https://devfeed.tech/topics/gemma.md>), [NeMo](<https://devfeed.tech/topics/nemo.md>)

Tags: [announcement](<https://devfeed.tech/tags/announcement.md>), [apache-2-0-tts](<https://devfeed.tech/tags/apache-2-0-tts.md>), [device](<https://devfeed.tech/tags/device.md>), [edge-tts-model](<https://devfeed.tech/tags/edge-tts-model.md>), [family](<https://devfeed.tech/tags/family.md>), [kleidiai](<https://devfeed.tech/tags/kleidiai.md>), [models](<https://devfeed.tech/tags/models.md>), [multilingual-tts](<https://devfeed.tech/tags/multilingual-tts.md>), [on-device](<https://devfeed.tech/tags/on-device.md>), [on-device-tts](<https://devfeed.tech/tags/on-device-tts.md>), [open](<https://devfeed.tech/tags/open.md>), [open-source-text-to-speech](<https://devfeed.tech/tags/open-source-text-to-speech.md>), [open-weight-tts](<https://devfeed.tech/tags/open-weight-tts.md>), [parameter](<https://devfeed.tech/tags/parameter.md>), [phonon](<https://devfeed.tech/tags/phonon.md>), [phonon-31m](<https://devfeed.tech/tags/phonon-31m.md>), [phonon-40m](<https://devfeed.tech/tags/phonon-40m.md>), [phonon-vs-kokoro](<https://devfeed.tech/tags/phonon-vs-kokoro.md>), [raspberry-pi-tts](<https://devfeed.tech/tags/raspberry-pi-tts.md>), [small-tts-model](<https://devfeed.tech/tags/small-tts-model.md>), [snapdragon-tts](<https://devfeed.tech/tags/snapdragon-tts.md>), [speech](<https://devfeed.tech/tags/speech.md>), [standard](<https://devfeed.tech/tags/standard.md>), [text](<https://devfeed.tech/tags/text.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [tts](<https://devfeed.tech/tags/tts.md>), [voice-cloning-on-device](<https://devfeed.tech/tags/voice-cloning-on-device.md>), [voice-design](<https://devfeed.tech/tags/voice-design.md>)

### AI overview

Phonon is an open-weight family of 31m and 40m parameter TTS models designed to synthesize speech locally, without a network connection. The 31m English model is 64 MB at q8 and Apache 2.0 licensed; the 40m models are 74.6 MB, cover English, French, German, Spanish, and Portuguese, and use CC BY-NC. The multilingual variant serves all five languages from one set of weights. The article reports 7 ms to first audio for Phonon 31m on a MacBook Air M5 CPU and 38 ms on a Raspberry Pi 5, where it runs at 3.7 times real time; results depend on device and benchmark setup. Models normalize structured text such as numbers and dates, support speed and loudness controls, and can use built-in or custom voices. The inference stack provides CPU, GPU, NPU, and browser backends. Developers can download weights and use the GitHub inference code, with integrations for Rust, Python, Swift, and JavaScript/WASM. Custom models are offered for tighter memory limits, additional languages, or product-specific vocabulary.

### Source excerpt

Today, we're open-sourcing Phonon, now a family of 31m and 40m parameter text-to-speech (TTS) models that set a new standard for naturalness and accuracy on-device.

## Gradium is now available on Eden AI

DevFeed: [Gradium is now available on Eden AI](<https://devfeed.tech/articles/gradium-is-now-available-on-eden-ai-90287.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/eden-ai-gradium-partnership>)

Author: Gradium

Published: 2026-10-19T00:00:00Z

Content type: article

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [litellm](<https://devfeed.tech/topics/litellm.md>), [asr](<https://devfeed.tech/topics/asr.md>), [Amazon Route 53](<https://devfeed.tech/topics/amazon-route-53.md>), [machine-translation](<https://devfeed.tech/topics/machine-translation.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-gateway](<https://devfeed.tech/tags/ai-gateway.md>), [ai-gateway-text-to-speech](<https://devfeed.tech/tags/ai-gateway-text-to-speech.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [api](<https://devfeed.tech/tags/api.md>), [data](<https://devfeed.tech/tags/data.md>), [data-residency](<https://devfeed.tech/tags/data-residency.md>), [eden](<https://devfeed.tech/tags/eden.md>), [eden-ai-gradium](<https://devfeed.tech/tags/eden-ai-gradium.md>), [eden-ai-tts-provider](<https://devfeed.tech/tags/eden-ai-tts-provider.md>), [eu](<https://devfeed.tech/tags/eu.md>), [eu-data-residency-voice-ai](<https://devfeed.tech/tags/eu-data-residency-voice-ai.md>), [european-text-to-speech-api](<https://devfeed.tech/tags/european-text-to-speech-api.md>), [gateway](<https://devfeed.tech/tags/gateway.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [gradium-tts](<https://devfeed.tech/tags/gradium-tts.md>), [llms](<https://devfeed.tech/tags/llms.md>), [models](<https://devfeed.tech/tags/models.md>), [region](<https://devfeed.tech/tags/region.md>), [routing](<https://devfeed.tech/tags/routing.md>), [speech](<https://devfeed.tech/tags/speech.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [speech-to-text-api](<https://devfeed.tech/tags/speech-to-text-api.md>), [speech-translation](<https://devfeed.tech/tags/speech-translation.md>), [speech-translation-api](<https://devfeed.tech/tags/speech-translation-api.md>), [text](<https://devfeed.tech/tags/text.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [translation](<https://devfeed.tech/tags/translation.md>), [vision](<https://devfeed.tech/tags/vision.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>)

### AI overview

Eden AI has added Gradium's text-to-speech, speech-to-text, and speech translation services to its API, with calls pinned to its EU region. The article explains that Gradium uses a passthrough integration to avoid the latency of a normalization layer and includes API examples and guidance on async transcription versus live streaming.

### Source excerpt

Eden AI is an AI gateway: one API for 500+ models across LLMs, speech, vision and translation, with routing, fallbacks and a dedicated EU endpoint for European data residency. Gradium's Text-to-Speech, Speech-to-Text and speech translation are now available through that API, pinned to the EU region.

## 1,000+ new voices, our most natural set yet

DevFeed: [1,000+ new voices, our most natural set yet](<https://devfeed.tech/articles/1-000-new-voices-our-most-natural-set-yet-81237.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/more-than-1000-voices>)

Author: Gradium

Published: 2026-10-07T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [AI for customer service automation](<https://devfeed.tech/topics/ai-for-customer-service-automation.md>)

Tags: [ai-voices](<https://devfeed.tech/tags/ai-voices.md>), [character-voices](<https://devfeed.tech/tags/character-voices.md>), [european-portuguese-tts](<https://devfeed.tech/tags/european-portuguese-tts.md>), [flagship-voices](<https://devfeed.tech/tags/flagship-voices.md>), [gradium-tts](<https://devfeed.tech/tags/gradium-tts.md>), [indian-english-tts](<https://devfeed.tech/tags/indian-english-tts.md>), [latin-american-spanish-tts](<https://devfeed.tech/tags/latin-american-spanish-tts.md>), [narration-voices](<https://devfeed.tech/tags/narration-voices.md>), [product](<https://devfeed.tech/tags/product.md>), [quebecois-french-tts](<https://devfeed.tech/tags/quebecois-french-tts.md>), [regional-accents-tts](<https://devfeed.tech/tags/regional-accents-tts.md>), [scottish-accent-tts](<https://devfeed.tech/tags/scottish-accent-tts.md>), [swiss-german-tts](<https://devfeed.tech/tags/swiss-german-tts.md>), [text-to-speech-accents](<https://devfeed.tech/tags/text-to-speech-accents.md>), [voice-agent-voices](<https://devfeed.tech/tags/voice-agent-voices.md>), [voice-catalog](<https://devfeed.tech/tags/voice-catalog.md>), [voice-design](<https://devfeed.tech/tags/voice-design.md>)

### AI overview

Gradium expanded its voice catalog with more than 1,000 voices across five languages and 28 accents, for customer service, narration, advertising, social media and character use. The release describes how automated scoring and native-speaker comparisons helped select the voices, and explains how to browse the catalog or use a voice ID in a text-to-speech request.

### Source excerpt

The Gradium voice catalog grows from ~380 to 1.4k voices. The 1,000+ new voices cover 28 accents across English, French, Spanish, German and Portuguese, cast for customer service, narration, ads and characters. Listen to 62 of them, and see how they were chosen.

## Gradium's TTS model targets sub-50 ms first audio with improved naturalness

DevFeed: [Gradium's TTS model targets sub-50 ms first audio with improved naturalness](<https://devfeed.tech/articles/why-we-built-the-fastest-robust-tts-model-81221.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/fastest-robust-tts-model>)

Author: Gradium

Published: 2026-10-01T00:00:00Z

Content type: article

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [asr](<https://devfeed.tech/topics/asr.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>)

Tags: [api](<https://devfeed.tech/tags/api.md>), [audio](<https://devfeed.tech/tags/audio.md>), [fastest-tts-model](<https://devfeed.tech/tags/fastest-tts-model.md>), [latency](<https://devfeed.tech/tags/latency.md>), [llm](<https://devfeed.tech/tags/llm.md>), [low-latency-tts](<https://devfeed.tech/tags/low-latency-tts.md>), [product](<https://devfeed.tech/tags/product.md>), [realtime](<https://devfeed.tech/tags/realtime.md>), [robust-tts](<https://devfeed.tech/tags/robust-tts.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [text-to-speech-latency](<https://devfeed.tech/tags/text-to-speech-latency.md>), [time-to-first-audio](<https://devfeed.tech/tags/time-to-first-audio.md>), [tts](<https://devfeed.tech/tags/tts.md>), [tts-benchmark](<https://devfeed.tech/tags/tts-benchmark.md>), [turn-gap](<https://devfeed.tech/tags/turn-gap.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>)

### AI overview

Gradium describes a text-to-speech model that starts delivering audio in under 50 ms and improves naturalness and expressivity while handling difficult inputs such as phone numbers and alphanumeric text. The article argues that lower latency leaves more of the conversational turn budget for the language model and handling interruptions. The model is available as the default for the API and studio, with no API changes required.

### Source excerpt

Our latest Text-to-Speech model takes under 50ms for the first audio chunk to arrive. Choosing a fast model does not have to come with cutting corners for model robustness: it improves naturalness and expressivity while retaining our strengths on hard cases like phone numbers and alphanumeric text.

## Giving voice agents a voice that adapts to context

DevFeed: [Giving voice agents a voice that adapts to context](<https://devfeed.tech/articles/giving-voice-agents-a-voice-that-adapts-to-context-81247.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/voice-that-adapts-to-context>)

Author: Gradium

Published: 2026-09-29T00:00:00Z

Content type: tutorial

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [livekit](<https://devfeed.tech/topics/livekit.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [article](<https://devfeed.tech/tags/article.md>), [code](<https://devfeed.tech/tags/code.md>), [context-aware-voice](<https://devfeed.tech/tags/context-aware-voice.md>), [customer-support-agents](<https://devfeed.tech/tags/customer-support-agents.md>), [decision-models](<https://devfeed.tech/tags/decision-models.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [programmable-voice](<https://devfeed.tech/tags/programmable-voice.md>), [prompt-to-voice](<https://devfeed.tech/tags/prompt-to-voice.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>), [voice-design](<https://devfeed.tech/tags/voice-design.md>), [voice-selection-pipeline](<https://devfeed.tech/tags/voice-selection-pipeline.md>)

### AI overview

A tutorial shows how to build a context-aware voice agent that classifies incoming messages, maps the signals to voice descriptions, and resolves requests to approved voices while creating missing voices in the background. It covers using a decision model, managing voice latency and voice libraries, and synthesizing replies with Gradium TTS.

### Source excerpt

Most voice agents speak with one preselected voice, whoever is on the line. Once a voice can be created from a text description, the voice becomes another property your code computes from context. This article builds the full pipeline: classify the message, map the signals to a Voice Design prompt, resolve it to an approved voice on the hot path, and design the missing ones in the background.

## How a benchmark change produced a faster TTS model

DevFeed: [How a benchmark change produced a faster TTS model](<https://devfeed.tech/articles/how-a-benchmark-change-produced-a-faster-tts-model-81217.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/coval-perceived-ttfa-benchmark>)

Author: Gradium

Published: 2026-09-09T00:00:00Z

Content type: article

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [AI-generated research reports](<https://devfeed.tech/topics/ai-generated-research-reports.md>)

Tags: [ai-benchmarks](<https://devfeed.tech/tags/ai-benchmarks.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [coval-tts-leaderboard](<https://devfeed.tech/tags/coval-tts-leaderboard.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [leading-silence](<https://devfeed.tech/tags/leading-silence.md>), [license](<https://devfeed.tech/tags/license.md>), [livekit](<https://devfeed.tech/tags/livekit.md>), [perceived-ttfa](<https://devfeed.tech/tags/perceived-ttfa.md>), [text-to-speech-latency](<https://devfeed.tech/tags/text-to-speech-latency.md>), [time-to-first-audio](<https://devfeed.tech/tags/time-to-first-audio.md>), [tts](<https://devfeed.tech/tags/tts.md>), [tts-benchmark](<https://devfeed.tech/tags/tts-benchmark.md>), [tts-benchmark-methodology](<https://devfeed.tech/tags/tts-benchmark-methodology.md>), [voice-agent-latency](<https://devfeed.tech/tags/voice-agent-latency.md>)

### AI overview

Coval changed its text-to-speech benchmark to include leading silence before the first audible sample, increasing Gradium's reported time to first audio despite no model or serving change. The article explains how the metric detects audible onset and why the new results cannot be compared directly with earlier leaderboard rows. Gradium says the measure prompted work on a model with no measured leading silence, which now reports a median TTFA of 214 ms.

### Source excerpt

On June 3, 2026, Gradium's time to first audio on the Coval TTS leaderboard moved from 171.9 ms to 429.6 ms. No model had shipped and no serving infrastructure had changed. What changed was the measurement: Coval had started counting the silence at the start of the stream.

## Gradium now available for free on LiveKit Inference

DevFeed: [Gradium now available for free on LiveKit Inference](<https://devfeed.tech/articles/gradium-now-available-for-free-on-livekit-inference-81236.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/livekit-gradium-partnership>)

Author: Gradium

Published: 2026-09-09T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [livekit](<https://devfeed.tech/topics/livekit.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [groq](<https://devfeed.tech/topics/groq.md>)

Tags: [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [api](<https://devfeed.tech/tags/api.md>), [bill](<https://devfeed.tech/tags/bill.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [conversational-ai](<https://devfeed.tech/tags/conversational-ai.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [gradium-tts](<https://devfeed.tech/tags/gradium-tts.md>), [inference](<https://devfeed.tech/tags/inference.md>), [latency](<https://devfeed.tech/tags/latency.md>), [livekit](<https://devfeed.tech/tags/livekit.md>), [livekit-gradium](<https://devfeed.tech/tags/livekit-gradium.md>), [livekit-inference](<https://devfeed.tech/tags/livekit-inference.md>), [livekit-tts-provider](<https://devfeed.tech/tags/livekit-tts-provider.md>), [low-latency](<https://devfeed.tech/tags/low-latency.md>), [low-latency-tts](<https://devfeed.tech/tags/low-latency-tts.md>), [robust](<https://devfeed.tech/tags/robust.md>), [speech](<https://devfeed.tech/tags/speech.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [text-to-speech-for-voice-agents](<https://devfeed.tech/tags/text-to-speech-for-voice-agents.md>), [ttfa](<https://devfeed.tech/tags/ttfa.md>), [tts](<https://devfeed.tech/tags/tts.md>), [use-cases](<https://devfeed.tech/tags/use-cases.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>), [voice-cloning](<https://devfeed.tech/tags/voice-cloning.md>)

### AI overview

Gradium Text-to-Speech is available through LiveKit Inference until Oct. 9 at no charge. The release describes support for voice cloning, conversational speech with low latency and robust pronunciation, and access through LiveKit without a separate Gradium account or API key. Gradium can be selected as the TTS provider for LiveKit voice agents, including through LiveKit's Agent Builder.

### Source excerpt

Gradium Text-to-Speech is now available on LiveKit Inference, free until Oct. 9. Gradium TTS is designed for conversational use cases, with voice cloning, low latency (below 250ms TTFA), and robust pronunciation on difficult cases.

## Launching Voice Design: prompt the voice your agent needs

DevFeed: [Launching Voice Design: prompt the voice your agent needs](<https://devfeed.tech/articles/launching-voice-design-prompt-the-voice-your-agent-needs-81235.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/launching-voice-design>)

Author: Gradium

Published: 2026-09-07T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [livekit](<https://devfeed.tech/topics/livekit.md>)

Tags: [ai-voice-generator](<https://devfeed.tech/tags/ai-voice-generator.md>), [apis](<https://devfeed.tech/tags/apis.md>), [custom-voices](<https://devfeed.tech/tags/custom-voices.md>), [product](<https://devfeed.tech/tags/product.md>), [prompt-to-voice](<https://devfeed.tech/tags/prompt-to-voice.md>), [regional-accents](<https://devfeed.tech/tags/regional-accents.md>), [synthetic-voices](<https://devfeed.tech/tags/synthetic-voices.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>), [voice-cloning-alternative](<https://devfeed.tech/tags/voice-cloning-alternative.md>), [voice-design](<https://devfeed.tech/tags/voice-design.md>)

### AI overview

Gradium's Voice Design feature creates synthetic voices from short text prompts for use in voice agents. The article describes its API and Studio workflow and presents a blind comparison across five languages, reporting that Gradium ranked first in each language among the tested systems.

### Source excerpt

Developers building voice agents ask us for voices matched to the use case in front of them: a Québécoise receptionist for a Montréal dealership, a Paulista support agent for a São Paulo fintech, a narrator in his sixties with the authority of a lecture hall. Briefs outnumber any catalog. Voice Design starts from a prompt: write one or two sentences, get complete new voices back in seconds, keep the one you want.

## Build voice agents with EU and US residency and Zero Data Retention

DevFeed: [Build voice agents with EU and US residency and Zero Data Retention](<https://devfeed.tech/articles/build-voice-agents-with-eu-and-us-residency-and-zero-data-retention-81219.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/eu-data-residency-zero-data-retention>)

Author: Gradium

Published: 2026-09-02T00:00:00Z

Content type: tutorial

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Zero Trust](<https://devfeed.tech/topics/zero-trust.md>), [zero trust network access](<https://devfeed.tech/topics/zero-trust-network-access.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [asr](<https://devfeed.tech/topics/asr.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Amazon Bedrock Knowledge Bases](<https://devfeed.tech/topics/amazon-bedrock-knowledge-bases.md>)

Tags: [ai-api](<https://devfeed.tech/tags/ai-api.md>), [amazon-sagemaker](<https://devfeed.tech/tags/amazon-sagemaker.md>), [api](<https://devfeed.tech/tags/api.md>), [audio-data-privacy](<https://devfeed.tech/tags/audio-data-privacy.md>), [aws](<https://devfeed.tech/tags/aws.md>), [data-processing-agreement](<https://devfeed.tech/tags/data-processing-agreement.md>), [data-residency](<https://devfeed.tech/tags/data-residency.md>), [data-residency-speech-api](<https://devfeed.tech/tags/data-residency-speech-api.md>), [data-retention](<https://devfeed.tech/tags/data-retention.md>), [eu-api-gradium-ai](<https://devfeed.tech/tags/eu-api-gradium-ai.md>), [eu-data-residency](<https://devfeed.tech/tags/eu-data-residency.md>), [eu-inference](<https://devfeed.tech/tags/eu-inference.md>), [eu-speech-api](<https://devfeed.tech/tags/eu-speech-api.md>), [europe](<https://devfeed.tech/tags/europe.md>), [gdpr-voice-ai](<https://devfeed.tech/tags/gdpr-voice-ai.md>), [gradium-security](<https://devfeed.tech/tags/gradium-security.md>), [product](<https://devfeed.tech/tags/product.md>), [speech-to-text-api-europe](<https://devfeed.tech/tags/speech-to-text-api-europe.md>), [text-to-speech-api-europe](<https://devfeed.tech/tags/text-to-speech-api-europe.md>), [us-api-gradium-ai](<https://devfeed.tech/tags/us-api-gradium-ai.md>), [us-data-residency](<https://devfeed.tech/tags/us-data-residency.md>), [voice-agent-compliance](<https://devfeed.tech/tags/voice-agent-compliance.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>), [x-gradium-residency](<https://devfeed.tech/tags/x-gradium-residency.md>), [zero-data-retention](<https://devfeed.tech/tags/zero-data-retention.md>)

### AI overview

The article explains how Gradium's data residency and Zero Data Retention settings work for voice inference. It describes pinning sessions to EU or US endpoints, rejecting requests that cannot stay in the selected region, checking residency status in API responses, and disabling payload retention. It also outlines self-hosted inference options through Amazon SageMaker or on-premises deployments.

### Source excerpt

Two questions come up in almost every enterprise evaluation: where exactly does the inference run, and what do you keep? Data residency pins your Text-to-Speech, Speech-to-Text and Speech-to-Speech sessions to our EU servers, or to our US servers, enforced at the server and reported on every response, so you can assert on it instead of trusting a hostname. Zero Data Retention decides whether the text and audio processed by our servers survive the call at all.

## Gradium releases a TTS model with 81% accuracy on hard cases and 216 ms median time to first audio

DevFeed: [Gradium releases a TTS model with 81% accuracy on hard cases and 216 ms median time to first audio](<https://devfeed.tech/articles/gradium-tts-you-no-longer-have-to-choose-between-latency-and-accuracy-81228.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/gradium-tts-latency-and-accuracy>)

Author: Gradium

Published: 2026-08-31T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [livekit](<https://devfeed.tech/topics/livekit.md>)

Tags: [best-tts-for-voice-agents](<https://devfeed.tech/tags/best-tts-for-voice-agents.md>), [cartesia-sonic-3-6](<https://devfeed.tech/tags/cartesia-sonic-3-6.md>), [coval-tts-benchmark](<https://devfeed.tech/tags/coval-tts-benchmark.md>), [elevenlabs-v3-conversational](<https://devfeed.tech/tags/elevenlabs-v3-conversational.md>), [email-address-text-to-speech](<https://devfeed.tech/tags/email-address-text-to-speech.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fish-audio-s2-1-pro](<https://devfeed.tech/tags/fish-audio-s2-1-pro.md>), [gradium-tts](<https://devfeed.tech/tags/gradium-tts.md>), [gradium-vs-cartesia](<https://devfeed.tech/tags/gradium-vs-cartesia.md>), [gradium-vs-elevenlabs](<https://devfeed.tech/tags/gradium-vs-elevenlabs.md>), [gradium-vs-fish-audio](<https://devfeed.tech/tags/gradium-vs-fish-audio.md>), [gradium-vs-inworld](<https://devfeed.tech/tags/gradium-vs-inworld.md>), [hard-case-tts-evaluation](<https://devfeed.tech/tags/hard-case-tts-evaluation.md>), [iban-text-to-speech](<https://devfeed.tech/tags/iban-text-to-speech.md>), [inworld-realtime-tts-2](<https://devfeed.tech/tags/inworld-realtime-tts-2.md>), [latency](<https://devfeed.tech/tags/latency.md>), [latency-accuracy-tradeoff](<https://devfeed.tech/tags/latency-accuracy-tradeoff.md>), [multilingual-text-to-speech](<https://devfeed.tech/tags/multilingual-text-to-speech.md>), [phone-number-text-to-speech](<https://devfeed.tech/tags/phone-number-text-to-speech.md>), [real-time-tts](<https://devfeed.tech/tags/real-time-tts.md>), [research](<https://devfeed.tech/tags/research.md>), [spelling-text-to-speech](<https://devfeed.tech/tags/spelling-text-to-speech.md>), [text-normalization-tts](<https://devfeed.tech/tags/text-normalization-tts.md>), [text-to-speech-accuracy](<https://devfeed.tech/tags/text-to-speech-accuracy.md>), [time-to-first-audio](<https://devfeed.tech/tags/time-to-first-audio.md>), [ttfa-p50](<https://devfeed.tech/tags/ttfa-p50.md>), [tts-accuracy-comparison](<https://devfeed.tech/tags/tts-accuracy-comparison.md>), [tts-hard-case-comparison-2026](<https://devfeed.tech/tags/tts-hard-case-comparison-2026.md>), [tts-latency](<https://devfeed.tech/tags/tts-latency.md>), [tts-pass-rate-benchmark](<https://devfeed.tech/tags/tts-pass-rate-benchmark.md>), [voice-agent-tts](<https://devfeed.tech/tags/voice-agent-tts.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>)

### AI overview

Gradium's new default text-to-speech model scores 81% on a 500-sentence evaluation of difficult spoken content across five languages. Its median time to first audio is 216 ms on Coval, with a 30 ms p75-p25 spread across 480 runs. Gradium says the model handles phone numbers, email addresses, IBANs, dates, and reference codes without preprocessing or text normalization; its evaluation dataset is publicly available.

### Source excerpt

A new Gradium TTS model is available today, and it is now the default. It reads the cases that break voice agents in production, phone numbers, email addresses, IBANs, with no pre-processing or text normalization on your side. Time to first audio (TTFA) P50 is 216 ms on Coval, 170 ms faster than the model it replaces, with a 30 ms p75 - p25 spread, the tightest of the five models tested. Here is the hard-case evaluation set, the head-to-head samples, and how to hear the difference yourself.

## Gradium voices are now live on AudioStack

DevFeed: [Gradium voices are now live on AudioStack](<https://devfeed.tech/articles/gradium-voices-are-now-live-on-audiostack-81216.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/audiostack-gradium-partnership>)

Author: Gradium

Published: 2026-08-19T00:00:00Z

Content type: news

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Holoscan for Media](<https://devfeed.tech/topics/holoscan-for-media.md>), [generative-ai-tools](<https://devfeed.tech/topics/generative-ai-tools.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-audio-production](<https://devfeed.tech/tags/agentic-audio-production.md>), [ai-voices-for-advertising](<https://devfeed.tech/tags/ai-voices-for-advertising.md>), [audio-production-at-scale](<https://devfeed.tech/tags/audio-production-at-scale.md>), [audiostack-gradium](<https://devfeed.tech/tags/audiostack-gradium.md>), [audiostack-voice-providers](<https://devfeed.tech/tags/audiostack-voice-providers.md>), [building](<https://devfeed.tech/tags/building.md>), [generation](<https://devfeed.tech/tags/generation.md>), [product](<https://devfeed.tech/tags/product.md>), [regional-accent-voices](<https://devfeed.tech/tags/regional-accent-voices.md>), [voice-ai-partnership](<https://devfeed.tech/tags/voice-ai-partnership.md>)

### AI overview

AudioStack has added Gradium as a voice provider. Gradium voices are available through AudioStack across multiple languages and regional accents, for an audio production platform that automates scripting, voice, music, sound design, mixing, and mastering.

### Source excerpt

AudioStack, the agentic audio production platform for media and advertising, has added Gradium to its roster of voice providers. Our voices are live on the platform now, across multiple languages and regional accents.

## New voices: how we pick the one that wins

DevFeed: [New voices: how we pick the one that wins](<https://devfeed.tech/articles/new-voices-how-we-pick-the-one-that-wins-81238.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/new-voices-how-we-pick-the-one-that-wins>)

Author: Gradium

Published: 2026-08-04T00:00:00Z

Content type: article

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Figma](<https://devfeed.tech/topics/figma.md>)

Tags: [accent-authenticity](<https://devfeed.tech/tags/accent-authenticity.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-voice-catalog](<https://devfeed.tech/tags/ai-voice-catalog.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [brand-voice](<https://devfeed.tech/tags/brand-voice.md>), [build](<https://devfeed.tech/tags/build.md>), [elo-benchmark](<https://devfeed.tech/tags/elo-benchmark.md>), [flagship-voices](<https://devfeed.tech/tags/flagship-voices.md>), [gradium-tts](<https://devfeed.tech/tags/gradium-tts.md>), [locale-specific-prompts](<https://devfeed.tech/tags/locale-specific-prompts.md>), [methodology](<https://devfeed.tech/tags/methodology.md>), [native-speaker-validation](<https://devfeed.tech/tags/native-speaker-validation.md>), [product](<https://devfeed.tech/tags/product.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [synthetic-voices](<https://devfeed.tech/tags/synthetic-voices.md>), [text-to-speech-voices](<https://devfeed.tech/tags/text-to-speech-voices.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-casting](<https://devfeed.tech/tags/voice-casting.md>), [voice-design](<https://devfeed.tech/tags/voice-design.md>), [voice-design-workshop](<https://devfeed.tech/tags/voice-design-workshop.md>), [voice-selection](<https://devfeed.tech/tags/voice-selection.md>)

### AI overview

Gradium describes how it selects flagship synthetic voices for particular use cases and locales. It generates and clusters hundreds of candidates, uses listener votes and head-to-head ELO rankings to shortlist them, and validates accent and pronunciation with native speakers. The article also explains why prompts need to be tailored to each locale and gives a five-step process for defining voice targets and evaluation scripts.

### Source excerpt

Picking a voice works like casting: one candidate in a thousand has a hit factor you cannot explain. Here is the methodology behind every flagship voice in the Gradium catalog, from generating hundreds of candidates per category to ranking them head-to-head against the incumbents on an ELO benchmark, why the same prompt does not travel across locales, and how to write the target yourself.

## Gradium TTS public beta handles complex text without pre-processing

DevFeed: [Gradium TTS public beta handles complex text without pre-processing](<https://devfeed.tech/articles/gradium-tts-public-beta-open-for-your-hardest-cases-81229.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/gradium-tts-public-beta>)

Author: Gradium

Published: 2026-07-30T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [vercel ai sdk](<https://devfeed.tech/topics/vercel-ai-sdk.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>)

Tags: [api](<https://devfeed.tech/tags/api.md>), [custom-voice-cloning](<https://devfeed.tech/tags/custom-voice-cloning.md>), [email-address-text-to-speech](<https://devfeed.tech/tags/email-address-text-to-speech.md>), [feedback](<https://devfeed.tech/tags/feedback.md>), [gradium-tts](<https://devfeed.tech/tags/gradium-tts.md>), [gradium-tts-beta](<https://devfeed.tech/tags/gradium-tts-beta.md>), [gradium-tts-public-beta](<https://devfeed.tech/tags/gradium-tts-public-beta.md>), [iban-text-to-speech](<https://devfeed.tech/tags/iban-text-to-speech.md>), [phone-number-text-to-speech](<https://devfeed.tech/tags/phone-number-text-to-speech.md>), [public-beta](<https://devfeed.tech/tags/public-beta.md>), [real-time-tts](<https://devfeed.tech/tags/real-time-tts.md>), [research](<https://devfeed.tech/tags/research.md>), [streaming-text-to-speech-api](<https://devfeed.tech/tags/streaming-text-to-speech-api.md>), [text-normalization-tts](<https://devfeed.tech/tags/text-normalization-tts.md>), [text-to-speech-beta](<https://devfeed.tech/tags/text-to-speech-beta.md>), [time-expressions-tts](<https://devfeed.tech/tags/time-expressions-tts.md>), [tts](<https://devfeed.tech/tags/tts.md>), [tts-pronunciation-accuracy](<https://devfeed.tech/tags/tts-pronunciation-accuracy.md>), [voice-agent-tts](<https://devfeed.tech/tags/voice-agent-tts.md>)

### AI overview

Gradium has released a public beta of its text-to-speech model that natively handles complex text such as phone numbers, email addresses, IBANs and time expressions without requiring developers to normalize the input. The model is available through the API, and beta participants who submit feedback receive 1 million Gradium credits.

### Source excerpt

A new Gradium TTS model is available today in public beta. It keeps improving on the current production model, handling the complex cases natively, phone numbers, email addresses, IBAN numbers, time expressions and more, with no pre-processing or text normalization on your side. Anyone can test it through the API, and everyone who sends us feedback during the beta receives 1M Gradium credits.

## Keyword boosting: teaching Speech-To-Text your vocabulary

DevFeed: [Keyword boosting: teaching Speech-To-Text your vocabulary](<https://devfeed.tech/articles/keyword-boosting-teaching-speech-to-text-your-vocabulary-81244.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/stt-keyword-boosting>)

Author: Gradium

Published: 2026-07-21T00:00:00Z

Content type: tutorial

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [asr](<https://devfeed.tech/topics/asr.md>), [voice ai](<https://devfeed.tech/topics/voice-ai.md>)

Tags: [asr](<https://devfeed.tech/tags/asr.md>), [asr-keyword-boosting](<https://devfeed.tech/tags/asr-keyword-boosting.md>), [boost-parameter](<https://devfeed.tech/tags/boost-parameter.md>), [custom-vocabulary-stt](<https://devfeed.tech/tags/custom-vocabulary-stt.md>), [domain-vocabulary-asr](<https://devfeed.tech/tags/domain-vocabulary-asr.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [gradium-stt](<https://devfeed.tech/tags/gradium-stt.md>), [keyword-boosting](<https://devfeed.tech/tags/keyword-boosting.md>), [proper-noun-transcription](<https://devfeed.tech/tags/proper-noun-transcription.md>), [semantic-vad](<https://devfeed.tech/tags/semantic-vad.md>), [speech](<https://devfeed.tech/tags/speech.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [streaming-speech-recognition](<https://devfeed.tech/tags/streaming-speech-recognition.md>), [stt](<https://devfeed.tech/tags/stt.md>), [stt-websocket-api](<https://devfeed.tech/tags/stt-websocket-api.md>), [voice-agent-transcription](<https://devfeed.tech/tags/voice-agent-transcription.md>), [word-error-rate](<https://devfeed.tech/tags/word-error-rate.md>)

### AI overview

This tutorial explains how keyword boosting helps speech-to-text models recognize rare, domain-specific terms without retraining. It recommends a boost value of 3 based on tests across 20 difficult clips, outlines dictionary rules for single-token keywords and case and accent variants, and shows how to configure the feature in Gradium's WebSocket API or studio.

### Source excerpt

Speech-To-Text models rarely see your product's vocabulary, brand names, drugs, athletes, and places, so they fall back on a similar-sounding common word. Keyword boosting raises the probability of the terms you care about while decoding, in real time, with no retraining. Here is how the boost parameter behaves, what a sweep across hard clips shows, and how to add it in a few lines.

## Gradium compares Phonon with on-device multilingual text-to-speech models

DevFeed: [Gradium compares Phonon with on-device multilingual text-to-speech models](<https://devfeed.tech/articles/smaller-clearer-better-phonon-multilingual-beats-magpie-and-neutts-nano-in-french-german-and-spanish-81240.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/phonon-multilingual-smaller-clearer-better>)

Author: Gradium

Published: 2026-07-15T00:00:00Z

Content type: comparison

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Local AI](<https://devfeed.tech/topics/local-ai.md>), [asr](<https://devfeed.tech/topics/asr.md>), [tokenization](<https://devfeed.tech/topics/tokenization.md>)

Tags: [100m-parameter-tts-model](<https://devfeed.tech/tags/100m-parameter-tts-model.md>), [beats](<https://devfeed.tech/tags/beats.md>), [better](<https://devfeed.tech/tags/better.md>), [edge-tts-model](<https://devfeed.tech/tags/edge-tts-model.md>), [english](<https://devfeed.tech/tags/english.md>), [error](<https://devfeed.tech/tags/error.md>), [errors](<https://devfeed.tech/tags/errors.md>), [fewer](<https://devfeed.tech/tags/fewer.md>), [french-tts](<https://devfeed.tech/tags/french-tts.md>), [german-tts](<https://devfeed.tech/tags/german-tts.md>), [magpie-tts](<https://devfeed.tech/tags/magpie-tts.md>), [model](<https://devfeed.tech/tags/model.md>), [multilingual-text-to-speech](<https://devfeed.tech/tags/multilingual-text-to-speech.md>), [multilingual-tts](<https://devfeed.tech/tags/multilingual-tts.md>), [neutts-nano](<https://devfeed.tech/tags/neutts-nano.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-magpie-tts](<https://devfeed.tech/tags/nvidia-magpie-tts.md>), [offline-tts](<https://devfeed.tech/tags/offline-tts.md>), [on-device](<https://devfeed.tech/tags/on-device.md>), [on-device-tts](<https://devfeed.tech/tags/on-device-tts.md>), [on-device-voice-cloning](<https://devfeed.tech/tags/on-device-voice-cloning.md>), [parameter](<https://devfeed.tech/tags/parameter.md>), [phonon](<https://devfeed.tech/tags/phonon.md>), [phonon-multilingual](<https://devfeed.tech/tags/phonon-multilingual.md>), [phonon-vs-magpie](<https://devfeed.tech/tags/phonon-vs-magpie.md>), [phonon-vs-neutts-nano](<https://devfeed.tech/tags/phonon-vs-neutts-nano.md>), [portuguese-tts](<https://devfeed.tech/tags/portuguese-tts.md>), [research](<https://devfeed.tech/tags/research.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [smallest-tts-model](<https://devfeed.tech/tags/smallest-tts-model.md>), [spanish-tts](<https://devfeed.tech/tags/spanish-tts.md>), [speaker](<https://devfeed.tech/tags/speaker.md>), [speaker-similarity](<https://devfeed.tech/tags/speaker-similarity.md>), [speech](<https://devfeed.tech/tags/speech.md>), [text](<https://devfeed.tech/tags/text.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [tts-voice-cloning](<https://devfeed.tech/tags/tts-voice-cloning.md>), [wavlm-large-speaker-similarity](<https://devfeed.tech/tags/wavlm-large-speaker-similarity.md>), [whisper-large-v3-tts-evaluation](<https://devfeed.tech/tags/whisper-large-v3-tts-evaluation.md>), [word-error-rate-tts](<https://devfeed.tech/tags/word-error-rate-tts.md>)

### AI overview

Gradium reports that its approximately 100M-parameter Phonon text-to-speech model outperforms several on-device voice-cloning models on word error rate and speaker similarity across tested languages. The comparison describes evaluation methods and notes that Magpie uses a fixed voice, so it has no comparable speaker-similarity score. The company says Phonon supports five languages and is designed for offline, latency-sensitive, and privacy-sensitive uses.

### Source excerpt

Phonon, our 100M-parameter on-device Text-To-Speech model, produces up to 3.5x fewer word errors than NVIDIA's Magpie TTS in French, German, and Spanish at 3.6x fewer parameters, and leads NeuTTS Nano on both word error rate and speaker similarity. It clones a reference voice across five languages: English, French, German, Spanish, and Portuguese.

## Natural conversation, live retrieval: Keenable search with Gradium

DevFeed: [Natural conversation, live retrieval: Keenable search with Gradium](<https://devfeed.tech/articles/natural-conversation-live-retrieval-keenable-search-with-gradium-81233.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/keenable-gradium-partnership>)

Author: Gradium

Published: 2026-07-10T00:00:00Z

Content type: news

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Omakub](<https://devfeed.tech/topics/omakub.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [Google Search](<https://devfeed.tech/topics/google-search.md>), [search-engines](<https://devfeed.tech/topics/search-engines.md>)

Tags: [announcement](<https://devfeed.tech/tags/announcement.md>), [conversational-search](<https://devfeed.tech/tags/conversational-search.md>), [gradbot](<https://devfeed.tech/tags/gradbot.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [keenable](<https://devfeed.tech/tags/keenable.md>), [live-retrieval](<https://devfeed.tech/tags/live-retrieval.md>), [low-latency-voice-ai](<https://devfeed.tech/tags/low-latency-voice-ai.md>), [open-source-voice-agent-framework](<https://devfeed.tech/tags/open-source-voice-agent-framework.md>), [realtime-web-search](<https://devfeed.tech/tags/realtime-web-search.md>), [retrieval-augmented-voice-agents](<https://devfeed.tech/tags/retrieval-augmented-voice-agents.md>), [voice-agent-latency](<https://devfeed.tech/tags/voice-agent-latency.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>), [voice-ai-search](<https://devfeed.tech/tags/voice-ai-search.md>), [web-retrieval-api](<https://devfeed.tech/tags/web-retrieval-api.md>)

### AI overview

Gradium and Keenable are partnering to add live web search to Gradbot, Gradium's open-source voice agent framework. The integration is intended to let voice agents retrieve fresh information during conversation while keeping retrieval fast enough to reduce pauses.

### Source excerpt

Gradium and Keenable are partnering to bring realtime web search into natural voice agents. Gradbot, Gradium's open source voice agent framework, is adding Keenable search so agents can retrieve fresh answers from the live web without breaking the flow of conversation.

## Gradium Extends Funding to $100 Million and Expands to Silicon Valley

DevFeed: [Gradium Extends Funding to $100 Million and Expands to Silicon Valley](<https://devfeed.tech/articles/gradium-extends-funding-to-100-million-and-expands-to-silicon-valley-81224.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/gradium-100-million-funding-nvidia>)

Author: Gradium

Published: 2026-07-08T00:00:00Z

Content type: news

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [Real-time AI interactions](<https://devfeed.tech/topics/real-time-ai-interactions.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [livekit](<https://devfeed.tech/topics/livekit.md>), [twilio programmable video](<https://devfeed.tech/topics/twilio-programmable-video.md>)

Tags: [agents-in-production](<https://devfeed.tech/tags/agents-in-production.md>), [ai-infrastructure](<https://devfeed.tech/tags/ai-infrastructure.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [conversational-ai](<https://devfeed.tech/tags/conversational-ai.md>), [funding](<https://devfeed.tech/tags/funding.md>), [generative-audio](<https://devfeed.tech/tags/generative-audio.md>), [gradium-100-million](<https://devfeed.tech/tags/gradium-100-million.md>), [gradium-funding](<https://devfeed.tech/tags/gradium-funding.md>), [gradium-nvidia](<https://devfeed.tech/tags/gradium-nvidia.md>), [gradium-nvidia-investment](<https://devfeed.tech/tags/gradium-nvidia-investment.md>), [gradium-san-francisco-office](<https://devfeed.tech/tags/gradium-san-francisco-office.md>), [gradium-seed-round](<https://devfeed.tech/tags/gradium-seed-round.md>), [kyutai](<https://devfeed.tech/tags/kyutai.md>), [neil-zeghidour](<https://devfeed.tech/tags/neil-zeghidour.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [real-time-voice-ai](<https://devfeed.tech/tags/real-time-voice-ai.md>), [san-francisco](<https://devfeed.tech/tags/san-francisco.md>), [seed-funding](<https://devfeed.tech/tags/seed-funding.md>), [silicon-valley-expansion](<https://devfeed.tech/tags/silicon-valley-expansion.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>), [voice-ai-funding](<https://devfeed.tech/tags/voice-ai-funding.md>), [voice-ai-infrastructure](<https://devfeed.tech/tags/voice-ai-infrastructure.md>), [voice-ai-startup-funding](<https://devfeed.tech/tags/voice-ai-startup-funding.md>)

### AI overview

Gradium announced an extension of its seed funding to $100 million, with NVIDIA among the new investors, and plans to establish a San Francisco Bay Area office. The company says the funding will support AI research, product development, international expansion, and its office. Its products include real-time speech-to-text, text-to-speech, speech translation, on-device speech generation, and an open-source framework for building voice agents.

### Source excerpt

Gradium extends its seed funding to $100 million, welcoming new investors including NVIDIA, and opens a San Francisco Bay Area office to scale its real-time voice AI.

## Gradium Powers RMC BFM Drive: AI-Generated Personalized Radio in Renault Vehicles

DevFeed: [Gradium Powers RMC BFM Drive: AI-Generated Personalized Radio in Renault Vehicles](<https://devfeed.tech/articles/gradium-powers-rmc-bfm-drive-ai-generated-personalized-radio-in-renault-vehicles-81242.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/rmc-bfm-drive-gradium>)

Author: Gradium

Published: 2026-07-07T00:00:00Z

Content type: news

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [voice ai](<https://devfeed.tech/topics/voice-ai.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-generated-radio](<https://devfeed.tech/tags/ai-generated-radio.md>), [ai-radio](<https://devfeed.tech/tags/ai-radio.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [broadcast-voice-synthesis](<https://devfeed.tech/tags/broadcast-voice-synthesis.md>), [connected-vehicles](<https://devfeed.tech/tags/connected-vehicles.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [in-car-voice-ai](<https://devfeed.tech/tags/in-car-voice-ai.md>), [openr-link](<https://devfeed.tech/tags/openr-link.md>), [personalized-radio](<https://devfeed.tech/tags/personalized-radio.md>), [pro-voice-clone](<https://devfeed.tech/tags/pro-voice-clone.md>), [real-time-text-to-speech](<https://devfeed.tech/tags/real-time-text-to-speech.md>), [renault](<https://devfeed.tech/tags/renault.md>), [rmc-bfm](<https://devfeed.tech/tags/rmc-bfm.md>), [rmc-bfm-drive](<https://devfeed.tech/tags/rmc-bfm-drive.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [voice-cloning](<https://devfeed.tech/tags/voice-cloning.md>)

### AI overview

Gradium powers RMC BFM Drive, a personalized radio service built into Renault connected vehicles. It uses AI to assemble programs from RMC BFM's content library based on drivers' interests, trip length, region, and preferred journalist. Gradium's text-to-speech models generate real-time transitions using cloned journalist voices.

### Source excerpt

Gradium powers RMC BFM Drive, the first fully AI-generated, personalized radio experience built into Renault connected vehicles. Our Text-To-Speech models synthesize transitions in real time using Pro Voice Clones of RMC BFM journalists, turning an AI-curated program into something that sounds and feels like a live broadcast.

## Gradium launches real-time speech translation models with selectable output voices

DevFeed: [Gradium launches real-time speech translation models with selectable output voices](<https://devfeed.tech/articles/launching-gradium-translate-the-best-accuracy-latency-tradeoff-against-gemini-3-5-live-translate-and-gpt-realtime-translate-81234.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/launching-gradium-translate>)

Author: Gradium

Published: 2026-06-24T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [machine-translation](<https://devfeed.tech/topics/machine-translation.md>), [livekit](<https://devfeed.tech/topics/livekit.md>), [communications](<https://devfeed.tech/topics/communications.md>)

Tags: [announcement](<https://devfeed.tech/tags/announcement.md>), [best](<https://devfeed.tech/tags/best.md>), [bleu](<https://devfeed.tech/tags/bleu.md>), [dubbing](<https://devfeed.tech/tags/dubbing.md>), [english](<https://devfeed.tech/tags/english.md>), [gemini](<https://devfeed.tech/tags/gemini.md>), [gemini-3-5-live-translate](<https://devfeed.tech/tags/gemini-3-5-live-translate.md>), [gpt-realtime-translate](<https://devfeed.tech/tags/gpt-realtime-translate.md>), [gradium](<https://devfeed.tech/tags/gradium.md>), [gradium-translate](<https://devfeed.tech/tags/gradium-translate.md>), [latency](<https://devfeed.tech/tags/latency.md>), [localization](<https://devfeed.tech/tags/localization.md>), [metricx](<https://devfeed.tech/tags/metricx.md>), [multilingual-voice-agents](<https://devfeed.tech/tags/multilingual-voice-agents.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [real-time-speech-to-text](<https://devfeed.tech/tags/real-time-speech-to-text.md>), [real-time-speech-translation](<https://devfeed.tech/tags/real-time-speech-translation.md>), [real-time-translation-api](<https://devfeed.tech/tags/real-time-translation-api.md>), [realtime](<https://devfeed.tech/tags/realtime.md>), [s2s-translate](<https://devfeed.tech/tags/s2s-translate.md>), [speech](<https://devfeed.tech/tags/speech.md>), [speech-to-speech](<https://devfeed.tech/tags/speech-to-speech.md>), [speech-to-speech-translation](<https://devfeed.tech/tags/speech-to-speech-translation.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [speech-to-text-translation](<https://devfeed.tech/tags/speech-to-text-translation.md>), [speech-translation](<https://devfeed.tech/tags/speech-translation.md>), [stt-translate](<https://devfeed.tech/tags/stt-translate.md>), [translation](<https://devfeed.tech/tags/translation.md>), [translation-benchmark](<https://devfeed.tech/tags/translation-benchmark.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-cloning-translation](<https://devfeed.tech/tags/voice-cloning-translation.md>), [voice-translation-api](<https://devfeed.tech/tags/voice-translation-api.md>)

### AI overview

Gradium announces stt-translate, which converts speech directly into translated text, and s2s-translate, which returns translated speech through a WebSocket. The models support five languages, and s2s-translate lets users choose a catalog voice or a clone. The post compares translation quality and latency with two competing services using BLEU and MetricX benchmarks.

### Source excerpt

Gradium launches stt-translate and s2s-translate: real-time Speech-To-Text and Speech-To-Speech translation across English, French, German, Spanish, and Portuguese, with a better accuracy-latency tradeoff than gpt-realtime-translate and gemini-3.5-live-translate and full control of the output voice, including cloning.

## Gradium TTS, upgraded: more accurate Text-To-Speech

DevFeed: [Gradium TTS, upgraded: more accurate Text-To-Speech](<https://devfeed.tech/articles/gradium-tts-upgraded-more-accurate-text-to-speech-81230.md>)

Original publisher: [Read original article](<https://gradium.ai/blog/gradium-tts-upgrade>)

Author: Gradium

Published: 2026-06-10T00:00:00Z

Content type: release

Language: en

Sources: [Gradium](<https://devfeed.tech/sources/gradium.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [Speech Processing](<https://devfeed.tech/topics/speech-processing.md>)

Tags: [agents-in-production](<https://devfeed.tech/tags/agents-in-production.md>), [api](<https://devfeed.tech/tags/api.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cartesia-sonic-3-5](<https://devfeed.tech/tags/cartesia-sonic-3-5.md>), [custom-voice-cloning](<https://devfeed.tech/tags/custom-voice-cloning.md>), [elevenlabs-flash-v2-5](<https://devfeed.tech/tags/elevenlabs-flash-v2-5.md>), [elevenlabs-multilingual-v2](<https://devfeed.tech/tags/elevenlabs-multilingual-v2.md>), [email-address-text-to-speech](<https://devfeed.tech/tags/email-address-text-to-speech.md>), [english](<https://devfeed.tech/tags/english.md>), [expressive-text-to-speech](<https://devfeed.tech/tags/expressive-text-to-speech.md>), [gradium-tts](<https://devfeed.tech/tags/gradium-tts.md>), [gradium-tts-upgrade](<https://devfeed.tech/tags/gradium-tts-upgrade.md>), [gradium-vs-cartesia](<https://devfeed.tech/tags/gradium-vs-cartesia.md>), [gradium-vs-elevenlabs](<https://devfeed.tech/tags/gradium-vs-elevenlabs.md>), [gradium-vs-inworld](<https://devfeed.tech/tags/gradium-vs-inworld.md>), [inworld-tts-1-5-max](<https://devfeed.tech/tags/inworld-tts-1-5-max.md>), [multilingual-text-to-speech](<https://devfeed.tech/tags/multilingual-text-to-speech.md>), [phone-number-text-to-speech](<https://devfeed.tech/tags/phone-number-text-to-speech.md>), [product](<https://devfeed.tech/tags/product.md>), [production](<https://devfeed.tech/tags/production.md>), [real-time-tts](<https://devfeed.tech/tags/real-time-tts.md>), [speech](<https://devfeed.tech/tags/speech.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [text-to-speech-model](<https://devfeed.tech/tags/text-to-speech-model.md>), [tts](<https://devfeed.tech/tags/tts.md>), [tts-acronyms](<https://devfeed.tech/tags/tts-acronyms.md>), [tts-benchmark-2026](<https://devfeed.tech/tags/tts-benchmark-2026.md>), [tts-pronunciation-accuracy](<https://devfeed.tech/tags/tts-pronunciation-accuracy.md>), [tts-prosody](<https://devfeed.tech/tags/tts-prosody.md>), [tts-spelling-accuracy](<https://devfeed.tech/tags/tts-spelling-accuracy.md>), [voice-agent-tts](<https://devfeed.tech/tags/voice-agent-tts.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>)

### AI overview

Gradium upgraded its text-to-speech model, reporting more natural speech and improved pronunciation of difficult inputs such as email addresses, phone numbers, acronyms, and codes. The company describes evaluation using objective metrics and listener tests, and says the model is now the default in its API with existing voices carried over.

### Source excerpt

Gradium TTS now runs on a new model: more natural prosody and substantially more accurate pronunciation on the cases that break voice agents in production, including spelling, acronyms, emails, phone numbers, and codes. It wins head-to-head against our previous model in all five languages, and leads real-time competitors (Cartesia Sonic 3.5, Inworld TTS 1.5 Max, ElevenLabs Flash v2.5 and Multilingual v2) on the hardest pronunciation cases. Available now as the new default, with custom voices carried over.

[Next page](<https://devfeed.tech/sources/gradium.md?cursor=WyIyMDI2LTA2LTEwVDAwOjAwOjAwKzAwOjAwIiwgIjgyNjdjODY3LTk1ZjAtNDU0Ni1hMDcyLTMxNGFkYTY3NWZlMiJd>)