Choosing Dograh vs ElevenLabs comes down to who runs the agent platform. ElevenLabs' ElevenAgents is a closed, hosted platform built around its own speech models. Dograh, an open-source, self-hostable AI voice agent platform, runs on your own servers or Dograh Cloud with any supported model, and ElevenLabs voices and Scribe speech-to-text run inside it.
Key Takeaways
- ElevenLabs voices, cloned ones included, run inside Dograh with your own key.
- ElevenAgents leads on keypad input, transfer types and burst concurrency.
- Dograh's code is open. ElevenLabs' private deployment is licensed through sales.
This post is part of our guide to ElevenLabs alternatives, where Dograh is the open-source platform that keeps ElevenLabs voices with your own key. Here we compare the two on license, hosting, models, builder and pricing, show how they work together and say who ElevenLabs suits.
The two platforms, row by row
ElevenLabs rents a hosted platform around its own voices, while Dograh is open-source software you can run anywhere, with any voice.
| Row | ElevenLabs (ElevenAgents) | Dograh |
|---|---|---|
| License | Closed-source platform with public SDKs | Open source under BSD-2-Clause, the whole platform |
| Where it runs | ElevenLabs' cloud, with U.S. data by default. EU, India or Singapore residency and private deployment in your own AWS or GCP account on Enterprise | Your own servers with Docker Compose, Dograh Cloud, or on-prem and private cloud on Enterprise. Self-hosting puts the servers on your team |
| How you pay | Plans from free (15 min) to $990/mo (12,375 min). Extra minutes $0.08, burst $0.16. Language model at the provider's rate, phone line billed by your carrier | Self-hosted: no per-minute platform fee. Dograh Cloud: a per-minute platform fee (see the last section). Models and phone billed by their providers |
| Models | ElevenLabs' own speech-to-text and voices. Hosted language models or your own OpenAI-compatible one. Realtime speech-to-speech, with no custom language model | Dograh's models, any supported provider with your own keys (ElevenLabs included), or open-weight models on your servers. Speech-to-speech from OpenAI, Google, Azure, Grok, Ultravox and AWS Nova Sonic |
| Builder | Visual Workflows with subagent, tool, transfer and end nodes | Visual workflow builder of nodes and pathways, plus an MCP server that lets an AI assistant create and edit agents |
| Phone | Native Twilio, SIP trunking and batch outbound calls | Your own carrier: Twilio, Vonage, Plivo, Telnyx, Cloudonix, Exotel, Vobiz or Asterisk ARI, or Dograh SIP. Campaigns for outbound lists |
| Transfer to a person | Conference (default), blind or SIP REFER | Transfer Call on Twilio, Telnyx, Plivo, Cloudonix and Asterisk ARI, with the details written to your system first |
| Keypad input | Reads callers' keypad digits and can press keys itself | No keypad (DTMF) tool. Callers answer by voice |
| Calls at once | 4 on Free up to 40 on Business, burst up to 3x or 300 at double the rate | Dograh Cloud includes 10, more on request. Self-hosted runs 10 by default, a limit you can raise |
| HIPAA | BAA on Enterprise only, with Zero Retention Mode for patient data | Self-host inside your own compliance perimeter, with no BAA with Dograh |
| Testing and QA | Automated tests, evals and A/B experiments | QA on test and live calls, scored per step, and call traces |
Licensed hosting and open source are different things. ElevenLabs' private deployment runs its code in your cloud account under its agreement, while Dograh's code is yours to read and change.
Using both: ElevenLabs voices inside Dograh
The two products also work together, which is the most useful fact in this comparison. ElevenLabs is a built-in provider in Dograh for both the voice and the transcription.
Open Providers in the sidebar, add an ElevenLabs connection with your API key, then create a configuration under Models in BYOK mode. On the voice tab, pick ElevenLabs, then a model from Flash v2.5, the default, up to Eleven v4 and v4 Turbo. Your ElevenLabs voices load from your key, and you can paste any voice ID by hand, cloned voices included. Scribe v2 Realtime is there for speech-to-text. The voice settings also let you point the base URL at an ElevenLabs Data Residency endpoint.
ElevenLabs bills those models to your own account at its API rates. Pre-recorded clips help here. Record the lines every call needs, and the agent plays a clip when one fits and falls back to a cloned ElevenLabs voice for the rest, so callers hear one voice throughout. Our post on cutting voice AI costs with pre-recorded audio works through what that saves.
Models and where the call audio goes
ElevenLabs picks the speech models inside an agent, while Dograh treats every model as your choice. In ElevenAgents, transcription, turn-taking and the voice come from ElevenLabs. You can bring your own language model through an OpenAI-compatible endpoint, though not with the Realtime speech-to-speech models.
Dograh is the orchestration layer that runs the call flow, tools and telephony, and the speech and language models are a separate layer you choose. Its OpenAI-compatible setting can point at a model on the same machine or network, such as one served by Ollama or vLLM.
Bringing your own key (BYOK) moves the contract, but the data still travels. With ElevenLabs in Dograh, the agent's replies go to ElevenLabs' API to become speech, and with Scribe the caller's audio goes there too, as with any provider. The audio stays in-house only when the models run on your own servers too, which our post on the open-source alternative to ElevenLabs walks through.
Open Source Alternative to Vapi / Retell
Self-hosted voice agent platform — no per-minute fees
dograh-hq/dograh
Star on GitHub
Phone, transfer and keypad input
ElevenLabs' call controls go further in two places, and both matter on support lines. ElevenLabs' transfer tool offers a conference transfer by default, a blind transfer and SIP REFER, and on native Twilio conference transfers it can pass a message to the person who picks up. ElevenAgents can also read the caller's keypad digits and press keys itself.
Dograh's Transfer Call works on Twilio, Telnyx, Plivo, Cloudonix and Asterisk ARI. Dograh calls the destination first, plays hold music to the caller while it rings, and connects the caller only when someone answers. On Twilio, it can play a short spoken introduction to both people before connecting them. No summary goes across by default, so an HTTP API tool writes the caller's details to your system before the transfer. If nobody answers within the 30-second default, the caller stays with the agent.
Dograh has no keypad (DTMF) tool, so callers answer by voice. On phone lines, ElevenAgents connects through Twilio or a SIP trunk you own, and Dograh connects to the carrier you already use, so either way your numbers stay in your carrier account.
Who should stay on ElevenLabs
ElevenLabs is the better fit for some teams, and the reasons are concrete. Stay if you want the top-ranked voices and the agent from one vendor on one bill, with no servers to run. Eleven v4 Turbo and Eleven v4 hold the top two places on the Artificial Analysis Speech Arena. Stay if your flow needs keypad input, a blind or SIP REFER transfer, or burst capacity for call spikes, or if your team relies on built-in tests, A/B experiments and SDKs for mobile and Unity.
Stay too if your buyers want a BAA from the vendor itself. ElevenLabs' HIPAA terms offer one on Enterprise, with Zero Retention Mode turned on for patient data. Dograh covers that need another way. Self-hosted inside your own compliance perimeter, it needs no BAA with Dograh. ElevenLabs needs its own, on Enterprise, if its voices handle patient data, and the phone carrier may need one depending on what it stores. Confirm the full setup with your counsel before patient calls go live. Our guide to HIPAA-compliant voice AI covers the full setup.
When Dograh is the stronger fit
Dograh is the stronger fit for teams that want to own the platform. Its code is open, so your security team can read what it approves, and you can change it or keep running your version whatever happens upstream. You pick every model, from ElevenLabs voices to local open-weight ones. QA runs on test and live calls and scores each step, and agents can be built from an AI assistant through Dograh's MCP server.
Self-hosting runs on Docker Compose and puts updates and uptime in your team's hands. Self-hosted Dograh has no per-minute platform fee (servers, model keys and phone minutes still cost); Dograh Cloud has a 1¢-a-minute platform fee. Our breakdown of ElevenLabs pricing covers what an ElevenAgents minute includes.
Run your real call script on both over a phone line, since most phone calls still reach the agent as 8 kHz audio while a browser demo runs at 16 kHz. To try Dograh, self-host it from GitHub, or start on Dograh Cloud.
Join the Dograh Community
Dograh is an OSS alternative to Vapi. Join our Slack community for queries, releases, best practices & community interactions.
Glossary
- ElevenAgents
- ElevenLabs' hosted voice agent platform, built around its own speech-to-text, turn-taking and voices. It bills by the call minute, with the language model and phone line on top.
- SIP REFER
- A SIP message that asks the phone network to move a call to another number, so the platform drops out of the call. ElevenAgents offers it as one of its transfer types.
- Zero Retention Mode
- ElevenLabs' setting that stops patient data from being stored in transcripts, recordings and logs. It is required, with an Enterprise BAA, before ElevenAgents handles protected health information.
- Bring your own key (BYOK)
- Using your own account with a model provider instead of the platform's. You hold the contract and the bill, but the call audio still goes to that provider.
