Cartesia voices

Your Sonic voice on a platform you run

Dograh is the self-hostable voice agent platform with an open-source workflow builder. Paste the same Sonic voice IDs your callers already know, clones too, and run them on open-source agents you host yourself or start on Dograh Cloud.

Book a Demo Self-host in Docker Try Now →

6,000+GitHub starsBSD-2open source$0platform fee, self-hostedAny systemconnects to the tools you use

01 / How it works

Key in, voice ID pasted, call placed

Example

Add your Cartesia keypaste a Sonic voice IDring the agent from a phone

01 · Connect

Cartesia is a built-in voice provider

BYOKbring your own key

Add your Cartesia key in the voice settings and choose Sonic 3.6, the default, or Sonic 3.5 or Sonic 3.

→ Cartesia bills the credits to your own account

02 · Voice

Any voice ID in your account

Voice IDthe string that names one voice

Paste a stock or cloned voice ID and set speed and volume beside it. Ink speech-to-text can hear the caller too.

→ Try ink-whisper first for a mixed-language line

03 · Blend

Real recordings, cloned fallback

Instant Clonea voice cloned from a short clip

Record your key lines, clone that speaker from about 10 seconds of audio, and the agent fills every other turn in the same voice.

→ Use the person who recorded the clips

02 / What fits together

Cartesia speaks, the rest is yours

From Cartesia

Sonic 3.6, 3.5 and 3Your cloned voice IDsInk speech-to-text

Cartesia still processes every line it voices, under your own contract

Runs the call

Visual workflow builderHTTP API tools mid-callTransfer Call to a person

Transfers run on Twilio, Telnyx, Plivo, Cloudonix and Asterisk ARI

Hosted where you choose

Your servers with DockerDograh CloudTwilio, Vonage, Plivo or Telnyx numbers

A self-hosted install starts at 10 calls at once, and you can raise it

03 / Why switch platforms

Change the platform, not the sound

Open code from builder to phone line

Read, fork and change the whole platform under BSD-2-Clause, with no sales call first.

Every model is your pick

Sonic can voice one agent while another uses one of 15 other voice providers, or Kokoro on your own hardware.

Move a number at a time

Your current agents keep answering while each rebuilt one passes a real phone test with the same voice ID.

No platform fee when self-hosted

Servers, model keys and phone minutes still cost. On Dograh Cloud the platform is 1¢ a minute.

04 / Dograh vs Managed Agents

What owning the platform adds

What you get	Dograh	Cartesia Managed Agents
Open-source code for the full agent platform	✓	✕
Install it yourself, no Enterprise sales deal	✓	✕
Open-weight language models on your own servers	✓	✕

05 / Guides

Cartesia alternatives, part by part

Complete guideCartesia Alternatives: Hosted Agent Platforms and Open SourceRead →ComparisonCartesia vs ElevenLabs: Voices, Speech-to-Text and AgentsCartesia vs ElevenLabs on the three layers a voice agent needs: Sonic 3.6 against Eleven v4 for voices, Ink against Scribe for speech-to-text, and Managed Agents against ElevenAgents for the platform, with list prices, hosting options, who each suits, and how both run inside Dograh with your own keys.ComparisonCartesia Features Explained: Sonic, Ink, Cloning and AgentsCartesia's features from its own docs: what Sonic 3.6 speaks and lets you control, how instant and professional voice cloning differ, what Ink 2 adds for turn detection, what Managed Agents does, and which of these settings you can use inside Dograh with your own key.ComparisonCartesia Pricing Explained: Credits, Agent Minutes and ExtrasCartesia pricing in plain numbers: monthly plans and credits, what Sonic voices and Ink speech-to-text use per character and per second, what a Managed Agents minute costs with the phone number and language model, and what Cartesia's voices cost when they run inside Dograh with your own key.AlternativesOpen-Source Alternative to Cartesia: Own the Agent, Pick the VoiceWhat is and is not open source at Cartesia, which open voices can stand in for Sonic, and how Dograh replaces Managed Agents as an open-source platform that keeps Cartesia's voices with your own key or runs open voices on your servers.

Questions

Can I use a cloned Cartesia voice?+

Yes. Paste the cloned voice ID in the voice settings with your own Cartesia key, and the agent speaks in that voice on every call.

Which Sonic models can I choose?+

Sonic 3.6 is the default, with Sonic 3.5 and Sonic 3 also offered. Speed and volume sit in the same settings.

Can Cartesia's Ink handle speech-to-text too?+

Yes. ink-whisper is the default and covers many languages, and Ink 2 is offered too, with Cartesia's own turn detection deciding when the caller has finished.

Does call audio still reach Cartesia?+

Yes. Every line Sonic voices, and with Ink the caller's audio, goes to Cartesia under your key. Open-weight voices keep audio on your servers.

Who pays for the Cartesia usage?+

Cartesia bills your own account in credits. Self-hosted Dograh adds no per-minute platform fee, and Dograh Cloud charges 1¢ a minute.

Can the agent hand a caller to a person?+

Yes, on Twilio, Telnyx, Plivo, Cloudonix and Asterisk ARI. The destination is rung first, and the caller joins once someone answers.

Paste your voice ID and ring the agent.

Self-host in one command, or start on Dograh Cloud.

Book a Demo Try Now →

$ curl … start_docker.sh && ./start_docker.sh