Voice & audio

Speech, transcription, and voices that were never in the room.

A best-for note is a starting point for a stated job. It is not a ranked bake-off. We did not run the products, and your laptop does not owe us a benchmark.

Latest briefs

Newest first. Each item has a five-minute take and the source.

Nov 5, 2026, 12:00 PM CST Filed

Watch

A speech model small enough to live on the device

Gradium's Phonon runs without a server, in about 65 MB and 75 MB sizes, across five languages. The English 31-million-parameter model is Apache 2.0. The other models are non-commercial. Local does not mean "sell it in your app."

Why it matters. The English 31-million-parameter model is Apache 2.0. The other models are non-commercial.

For people who want on-device speech. Recheck the model card before you ship a language other than that English model.

Voice & audio Source: Hugging Face

Corrections: if a fact is wrong, email hello@loadedtokens.com.

Nov 2, 2026, 12:00 PM CST Filed

Watch

GPT-Live can place the phone call

Outbound SIP calling is now available to every organization on the Live API. You connect GPT-Live to a number through your own SIP provider. The changelog is the availability note. The price sits on a separate pricing page. You still bring the phone plan.

Why it matters. Outbound calling is available on the Live API, and you still bring the phone plan. The price is on a separate page.

For organizations that already have a SIP provider. Recheck the pricing page before you budget calls.

Voice & audio Source: OpenAI API changelog

Corrections: if a fact is wrong, email hello@loadedtokens.com.

Best for

UseTranscribe audio you send to an API

Deepgram

Publishes per-minute speech-to-text rates and a free credit. You send the audio. They send the transcript.

Read the profile →

UseTranscribe on your own machine

Whisper

The local speech model. Downloading it is not the same as calling OpenAI's transcription API.

Read the profile →

WatchA hosted voice in a product

ElevenLabs Text to Speech

A hosted text-to-speech service, for when the voice has to live inside your product.

Read the profile →

Tool verdicts

Use

AssemblyAI

Speech-to-text with a published hourly rate, plus text to speech.

$50 in free credits. Universal-3.5 Pro pre-recorded is $0.21 an hour. Streaming bills the open socket, not just the audio. Text to speech is $15 per million characters.

Watch

Cartesia

Speech models and voice agents sold as credits.

Free is 20,000 credits a month. A $49 plan includes 1.25 million credits. Voice-agent minutes are a separate prepaid balance.

Use

Deepgram

Speech-to-text, text-to-speech, and voice agents.

$200 free credit, then pay as you go. Nova-3 monolingual pre-recorded is $0.0043 a minute. Aura-2 is $0.030 per 1,000 characters.

Watch

ElevenLabs Scribe

Speech-to-text tools for recorded and real-time transcription, with timestamps and speaker features.

Free and paid shared-credit plans. The speech-to-text page lists plan-specific minutes; Starter is $6/month before taxes after promotion.

Watch

ElevenLabs Text to Speech

Speech generation with plan-dependent voice features, credit usage and licensing.

Free $0/month with 10,000 shared credits; Starter $6/month after first-month promotions. Commercial license listed on Starter.

Watch

GPT-Live

OpenAI's live voice session, including outbound SIP.

gpt-live-1 is $0.05 a minute, billed per second. Backend model and tool use are charged separately. Realtime audio tokens are a different table, at $32 in and $64 out per million for gpt-realtime-2.1 audio.

Use

Whisper

OpenAI's open speech-recognition models and code.

The repository is free to download. OpenAI's hosted transcription API is a separate, metered product.

All resources · Full directory · This category as JSON

Updates that matter in five minutes or less.

Free, Tuesdays & Fridays.