ElevenLabs Pricing Explained: Agent Plans, Minutes and Extras

ElevenLabs prices ElevenAgents by the call minute. Plans run from free with 15 minutes to $990 a month for 12,375 minutes, and every paid plan works out to $0.08 a minute, with the language model and phone line on top. ElevenLabs voices also run in Dograh, an open-source AI voice agent platform you can self-host or run on Dograh Cloud.

Key Takeaways

• Every paid plan works out to $0.08 per included minute.
• The language model and the phone line are billed on top of that minute.
• Data residency and private deployment come only with an Enterprise contract.

This post is part of our guide to ElevenLabs alternatives, where ElevenAgents bills by the call minute and the language model and phone line come on top. Here we take the plans, each charge outside the minute, a worked month, the API prices and what ElevenLabs voices cost inside Dograh.

ElevenAgents plans and what a minute costs

ElevenAgents sells monthly plans with included call minutes, and the plan you pick mostly decides how many calls can run at once. ElevenLabs' agent pricing lists seven tiers. Agent plans are billed by call minutes, separately from the credit pool ElevenLabs' creator plans use.

Divide any paid plan's price by its minutes and you get $0.08 a minute, by our arithmetic. Pro's $99 for 1,238 minutes is $0.080, and Business's $990 for 12,375 is the same. Extra minutes also cost $0.08, so a bigger plan buys more calls at once, seats and voice clones while the minute price stays put. Enterprise adds "significant discounts at scale", priced through sales.

Text messages in an agent conversation cost $0.003 each. Creating an agent costs nothing.

What sits on top of the $0.08

The $0.08 covers ElevenLabs' own speech models in one agent pipeline, and three things are billed beside it. The first is the language model, billed as an LLM pass-through. ElevenLabs charges the provider's published rate with no markup, except some Gemini and Claude models, which cost 10% more to match Google's regional Vertex rates. These are priced per million tokens, so the cost per minute depends on how much the agent says and reads. Each agent's model picker shows its own estimate under "Detailed costs".

The second is the phone line. ElevenLabs charges nothing for telephony, so your Twilio account or SIP carrier bills you directly. On Twilio's US voice pricing, receiving a call on a local number costs $0.0085 a minute, plus $1.15 a month for the number. A toll-free number costs $0.022 a minute to receive calls, plus $2.15 a month.

The third is burst capacity. ElevenLabs' burst pricing lets an agent take up to three times its plan's concurrency, or 300 calls, whichever is lower, at $0.16 a minute for the calls above the limit. It is switched on per agent.

Two billing rules change how minutes are counted. ElevenLabs' cost FAQ gives a 95% discount on silence longer than 10 seconds, and it bills on connection time, so you may pay for slightly more than the audible call. Realtime speech-to-speech agents cost the same $0.08 a minute plus tokens, $0.20 in and $6.00 out per million on Realtime v1 Mini, or $1.80 and $9.00 on Max.

Open Source Alternative to Vapi / Retell

Self-hosted voice agent platform — no per-minute fees

dograh-hq/dograh

Star on GitHub

A worked month

Here are two inbound months on Twilio local numbers, worked out by our own arithmetic from the prices above. Because every plan's minutes cost the same, pick the plan by the calls you need at once, then pay $0.08 for the rest.

Creator allows 10 calls at once and Pro 20. If your line needs more, the same 3,000 minutes on Scale is $299, with 30 calls at once. The language model comes on top of every total, and a toll-free line moves the phone row to $0.022 a minute.

ElevenLabs API prices and Enterprise-only costs

ElevenLabs also sells its speech models on their own, which is what you pay when its voices run inside another platform. ElevenLabs' API pricing charges per 1,000 characters for text to speech, $0.04 on Flash, Turbo, v3 Conversational and v4 Turbo, and $0.08 on Multilingual v2, v3 and v4. Speech-to-text costs $0.22 an hour on Scribe v2 and $0.39 an hour on Scribe v2 Realtime, about $0.0065 a minute.

Some needs sit on Enterprise only. A business associate agreement (BAA) under the Health Insurance Portability and Accountability Act (HIPAA), which also needs Zero Retention Mode, comes only there. So does EU, India or Singapore data residency, and private deployment in your own AWS or GCP account is arranged through sales. A team with any of those needs prices ElevenAgents through a contract instead of the plan table.

ElevenLabs voices inside Dograh

ElevenLabs voices are a built-in provider in Dograh, used with your own key, so ElevenLabs bills them at the API rates above. The voice settings cover Flash v2.5 up to Eleven v4 Turbo, cloned voice IDs included, and Scribe v2 Realtime for speech-to-text.

Here is one minute with those models, by our own arithmetic. At about 400 spoken characters a minute, the same assumption as our Vapi pricing breakdown, Flash at $0.04 per 1,000 characters is about $0.016 for the voice. Scribe v2 Realtime adds about $0.0065. The language model and the phone line come on top, as with ElevenAgents. 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. Dograh's own managed models, if you use them instead, are about $0.07 a minute for most models, billed apart from that fee.

These lines cover different layers, so they are not a like-for-like price against ElevenAgents' $0.08, which also runs the turn-taking and hosts the agent. Pre-recorded audio can cut the voice line further, because clips play directly and are not billed as text to speech. Our post on cutting voice AI costs with pre-recorded audio works through how much that saves.

When ElevenLabs' pricing fits

ElevenLabs' pricing suits teams that want to start small and keep the agent with one vendor. The free plan gives 15 minutes to test, Starter costs $6 a month, and voices, speech-to-text and turn-taking arrive on one bill with no servers to run. The flat $0.08 makes a month easy to forecast once you know your minutes.

The plan table stops fitting at the edges. Steady traffic above your concurrency limit pays double through burst, so a larger plan often costs less than daily bursting. A BAA, data residency or private deployment moves you to an Enterprise contract. Our Dograh vs ElevenLabs comparison covers the features behind those lines, row by row.