ElevenLabs is the voice layer in a short-form stack. A faceless YouTube channel, a TikTok that uses a cloned narrator, and a Shorts editor who dubs an existing cut into another language all hit the same credit pool. The product also sells speech to text, sound effects, music, voice isolation, and a studio project timeline, and it sells ElevenAgents for conversational voice. This directory lists it with video tools because the voice is what makes a generated picture publishable. It does not generate the picture. Pair it with Seedance, Kling, Runway, or CapCut when the frame still has to be made or cut.

The prices below come from elevenlabs.io/pricing, opened on 2026-10-16 on the monthly billing view. A separate API price page at elevenlabs.io/pricing/api was not observed in this check. Do not mix an older "characters per dollar" screenshot into these plan stickers. Where the creative plan mentions the API, this page quotes that mention and stops.

Who it's for

ElevenLabs fits anyone whose bottleneck is the voice track. A creator who films or generates silent b-roll and needs a consistent narrator uses text to speech or an instant clone. A creator who already has a voice and wants it in another language uses dubbing. A team that shares a workspace uses Scale or Business, which are the plans that print seat counts. A developer who needs 44.1 kHz PCM from the API is pushed toward Pro by the plan card, which lists that output as a Pro feature.

Free is a real $0 plan with 10,000 credits a month, and the FAQ says rollover does not apply to it. Starter adds a commercial license and instant voice cloning. Creator adds professional voice cloning. People who only need to hear a draft can stay on Free. People who publish the audio under a brand should read the commercial-license line before they ship a Free export. This check did not open the license text, so the exact commercial grant is the plan card's label plus the terms, not a paraphrase invented here.

It is the wrong primary tool when the job is an avatar face. HeyGen and Creatify include voices because their product is a person on screen. Use ElevenLabs when you will lay the audio into an editor yourself, or when HeyGen's included voice is not the one you want. It is also the wrong tool for burned-in captions. CapCut remains the caption bench. ElevenLabs can supply the track those captions follow.

What it does

The pricing page groups the product into text to speech, speech to text, sound effects, voice design, music, productions, image, and studio projects. Starter adds a commercial license, instant voice cloning, music commercial use, dubbing studio, and image and video. Creator adds professional voice cloning. Pro adds 44.1 kHz PCM via the API and 192 kbps audio. Scale adds 3 workspace seats, team collaboration, and 3 professional voice clones. Business adds low-latency text to speech "as low as 5c/minute," 10 professional voice clones, and 10 seats. Enterprise is custom, with SSO, BAAs for HIPAA customers, and more seats and voices.

The compare table, flattened in reading order across Free, Starter, Creator, Pro, Scale, and Business, showed about 10, 30, 121, 600, 1,800, and 6,000 minutes included. Those figures match the credit grants if you treat the page's approximate text-to-speech cost as the converter, and they are printed with a tilde, so they are estimates. Extra-minute prices on the same table were about $0.36, $0.20, $0.18, $0.17, $0.17, and $0.17 in that plan order. Languages were listed as 74 on each of those columns. Custom voice slots read as 3, 10, 30, 160, 660, and 2,200. Concurrent requests read as 2, 3, 5, 10, 15, and 25. Audio quality read as 128 kbps at 44.1 kHz on Free and Starter, and 128 and 192 kbps at 44.1 kHz from Creator upward. Quote these as compare-table figures observed on 2026-10-16, still marked approximate where the page used a tilde.

Credits are shared. The FAQ says the same monthly pool pays for every product, so a music bed and a voiceover compete. Approximate costs on that FAQ: text to speech 1 credit per character for the V2 multilingual models; speech to text 330 credits per minute; music 900 credits per minute; sound effects 200 credits per generation; voice changer and voice isolator 1,000 credits per minute; dubbing 2,000 credits per minute for automatic with watermark, 3,000 without watermark, 5,000 for Dubbing Studio with watermark, and 10,000 without watermark. Flash and Turbo models can cost 0.5 to 1 credit per character on the API. A single request has a maximum credit count. The numeric cap was not observed. It depends on free versus paid.

Pricing tiers

Verified on 2026-10-16. Prices exclude taxes. The Starter card said "Until Oct 18" over a first month at $1 (83 percent off the $6 price). Creator's first month is $11, which the FAQ also states. This check is 16 October 2026, so that Starter date was still in the future on the day the page was read. It may be gone if you open the page later. The renewal prices are the ones to budget.

Plan	Monthly price	Credits	Seats and extras printed on the card
Free	$0	10,000 / month	3 studio projects. No rollover
Starter	$6, first month $1 until Oct 18	30,000 / month	Commercial license, instant voice cloning, 20 studio projects
Creator	$22, first month $11	121,000 / month	Professional voice cloning, additional credits
Pro	$99	600,000 / month	44.1 kHz PCM via API, 192 kbps
Scale	$299	1,800,000 / month	3 seats, 3 professional voice clones
Business	$990	6,000,000 / month	10 seats, 10 professional voice clones, low-latency TTS as low as 5 cents/minute
Enterprise	Custom	Custom	Custom terms, SSO, more seats and voices

Annual billing, from the FAQ: you pay for 10 months, which the page calls two months free. The equivalent monthly price is $5 for Starter, $18.33 for Creator, $82.50 for Pro, $249.17 for Scale, and $825 for Business. Those are the FAQ's figures, not a separate checkout this directory completed.

Rollover: unused credits on a paid plan roll for up to two months, up to 2 times the monthly quota, so the balance can reach at most 3 times the monthly quota while you stay on that paid plan and do not downgrade or cancel. Downgrading or canceling forfeits unused credits at the end of the cycle. Free does not roll. Pay-as-you-go top-up credits are separate and are not under that cap. The dollar price of a top-up pack was not observed on this creative pricing page.

A worked example, using only the FAQ's 1 credit per character for V2 multilingual text to speech: 121,000 Creator credits are on the order of 121,000 characters, which the compare table calls about 121 minutes. A 30-second Shorts narration is a small slice of that grant. Dubbing is not. Automatic dubbing without a watermark at 3,000 credits per minute spends a Creator month in about 40 minutes of source audio. Price the mode you will actually run.

Per-1,000-character API dollar rates on the dedicated API pricing page were not observed in this check. Pro's card is the verified API-related fact: 44.1 kHz PCM output via the API is listed from Pro up. Runway also resells some ElevenLabs audio models inside Runway's own API credits. That is Runway's price, not this subscription.

Pros

The monthly stickers and the credit grants are both on the official pricing page, including the first-month discounts and the date on the Starter offer.
One credit pool is documented, with approximate per-product costs, so a creator can see that dubbing is an order of magnitude more expensive than a short narration.
Annual math is printed as monthly price times 10, with the equivalent monthly numbers written out.
Paid rollover lasts up to two months, which is explicit, and the free-plan exception is equally explicit.
The API reference is linked from the pricing page, and an official hosted MCP server publishes a URL, an OAuth flow, and regional endpoints.

Cons

The first-month prices end. Starter's $1 is labeled until 18 October. Creator's $11 is a first month. The $6 and $22 renewals are the plan.
Character costs depend on the model. V2 multilingual at 1 credit per character is not the Flash or Turbo rate, and it may not be the v3 or v4 rate. The FAQ's "about 1 credit per character" for text to speech is an approximation.
The dedicated API dollar table was not opened. Anyone budgeting a high-volume API integration still needs that page.
Commercial terms, the single-request credit cap, and the price of extra credits were not fully specified in the extract.
ElevenLabs does not finish the picture. A voice-only subscription still needs an editor and, for faceless video, a generator.
The hosted MCP server is not a dump of every API. Dubbing, music, and speech to text were not in the MCP tool list that was read.

Usage flow

Pick the renewal price, not the first-month price. Use Free to hear a voice. Use Starter or above when the commercial-license line matters. Use Creator or above when you need a professional clone.
Generate the narration, the dub, or the sound effect, and read the credit estimate before you run a Dubbing Studio job. A watermarked automatic dub and a clean Dubbing Studio job are not the same credit cost.
Lay the file into CapCut or the avatar tool, and burn captions there. If the face must move with the new language, compare HeyGen, which sells lip-sync translation as its own product.

Workflow usage

On the voice-publish workflow (voice-publish), ElevenLabs is the primary tool. A channel records or writes, generates or clones the voice, and exports a track that other tools do not replace. The Starter commercial license is the line to confirm before that track goes on a monetized Short. Scale and Business matter when more than one editor must share the clone. Seat counts are 3 and 10 on those cards.

On the faceless workflow (faceless), ElevenLabs is the narration step after a model such as Kling or Seedance has made the shots. A 121,000-credit Creator pool is large relative to a few Shorts scripts and small relative to dubbed long episodes. Do the dubbing math before you promise a localized back catalog. This site does not yet serve live pages at /workflow/voice-publish or /workflow/faceless. The slugs are the planning labels.

Underlying models

The pricing page names Eleven v4 as a current emotive model in the page header, and the compare table lists Eleven v4 and Eleven v3 as rows. The FAQ discusses V2 multilingual at 1 credit per character and discounted Flash and Turbo models. A full credit table for v3 and v4 was not observed on the creative pricing FAQ. Treat "1 credit per character" as the V2 multilingual figure the FAQ states, not as a universal rate. Speech-to-text, music, and dubbing have their own approximate rates above, and those rates are about the product, not about a single named checkpoint.

API and MCP

The pricing page links an API reference at https://elevenlabs.io/docs/api-reference/text-to-speech/convert. That URL is the api_url on this entry because it is the official text-to-speech reference the pricing page points at. The request schema and the per-character dollar price on elevenlabs.io/pricing/api were not observed on 2026-10-16. has_api is true because the company publishes that API and sells API audio quality on the Pro card.

The official hosted MCP server is documented at elevenlabs.io/docs/eleven-agents/operate/hosted-mcp, checked the same day. The default URL is https://api.elevenlabs.io/v1/mcp. Authentication is OAuth. No API key is copied into the client. Regional residency URLs on that page are https://api.eu.residency.elevenlabs.io/v1/mcp, https://api.in.residency.elevenlabs.io/v1/mcp, and https://api.sg.residency.elevenlabs.io/v1/mcp. The documented actions are agent management — create and update agents, read transcripts, manage tools and knowledge — and generating speech audio from text as a short-lived download link. Access is scoped to the OAuth grants you approve, covering ElevenAgents and text to speech. It is not a documented dubbing, music, or speech-to-text MCP. A third-party local MCP package is not this server. Point clients at the hosted URL.

Similar tools

HeyGen includes voices because the product is an avatar. Its MCP spends HeyGen web-plan credits and is the right server when the agent must render a face.
Creatify bundles text to speech inside an ad, at 1 Creatify credit per 30 seconds on the API card. That is a different meter.
CapCut is where the ElevenLabs track becomes a captioned Short. CapCut's own premium text to speech is a separate credit, described on CapCut's help pages, and it does not replace a cloned ElevenLabs voice.
Runway can call selected ElevenLabs audio models on Runway API credits. Use that only when the rest of the job is already a Runway API call.

FAQ

Is the Starter plan $1 or $6?

Both, on the page checked 2026-10-16. The card showed $1 for the first month, labeled until 18 October, and $6 as the plan price. The FAQ lists Starter at $6 for 30,000 credits. Budget $6 after the first month. Creator is $11 for the first month and $22 after, for 121,000 credits.

Do credits roll over?

On paid plans, yes, for up to two months, capped so the balance stays at most three times the monthly grant, and only while you do not downgrade or cancel. Free credits do not roll over. Top-up credits are a separate balance. The price of those top-ups was not on the creative pricing extract.

What does the official MCP server do?

It manages ElevenAgents and can generate speech audio as a short-lived download link. The URL is https://api.elevenlabs.io/v1/mcp, with EU, India, and Singapore residency hosts. It uses OAuth. It is not a substitute for the dubbing or music APIs.

How many credits does a dubbed minute cost?

The pricing FAQ says about 2,000 credits per minute for automatic dubbing with a watermark, 3,000 without, 5,000 for Dubbing Studio with a watermark, and 10,000 for Dubbing Studio without a watermark. Text to speech is a different rate, about 1 credit per character on V2 multilingual models.