# ElevenLabs vs Vapi: Cost per Call and a Launch Checklist > ElevenLabs Agents vs Vapi for a small business phone receptionist: cost per call worked out, setup effort compared, and a launch checklist. Source: https://learn.ynteractive.com/content/elevenlabs-vs-vapi-cost-per-call-launch-checklist · Updated 2026-10-10 · Free from Stepthrough (https://learn.ynteractive.com) ## Practice it step by step [Start the free practice](https://learn.ynteractive.com/modules/voice-agents-compare-launch) You do every step yourself in a practice copy of ChatGPT. It checks your work as you go, and “Do it for me” finishes any step you get stuck on. Free, about 8 minutes, and nothing touches your real accounts. [Open full screen](https://learn.ynteractive.com/modules/voice-agents-compare-launch) Free · Live now · [AI Voice Agents](https://learn.ynteractive.com/content/course/ai-voice-agents) · [ChatGPT](https://learn.ynteractive.com/content/tool/chatgpt) · ElevenLabs · Vapi · ~8 min · 19 steps · Beginner Updated October 10, 2026. Built against the tool’s current screens and checked step by step. [How we check this lesson](https://learn.ynteractive.com/content/about#how-we-check) [ElevenLabs Agents](https://learn.ynteractive.com/content/tool/elevenlabs-agents) and Vapi can both run an [AI receptionist](https://learn.ynteractive.com/content/glossary/ai-receptionist) that answers your phone after hours. Which one costs less, and which one is easier to live with? In this free module you take the same receptionist built on both platforms earlier in the course, give [ChatGPT](https://learn.ynteractive.com/content/tool/chatgpt) the prices from each pricing page and your call volume, work out the monthly cost and the cost per call (and catch a math slip), compare the setup effort, and finish with a launch checklist and an honest greeting. ## At a glance | Question | Answer | | --- | --- | | What you build | Decide between the ElevenLabs and Vapi versions of Cedar Creek Plumbing's AI receptionist: give ChatGPT the real prices, catch its math slip, compare volumes and setup effort, then build a launch checklist and an honest greeting. | | Tools | ChatGPT | | Time | About 8 minutes, 19 steps in 6 parts | | Level | Beginner, no coding needed | | Cost | Free | | Ways to learn it | [Practice it step by step](https://learn.ynteractive.com/modules/voice-agents-compare-launch) | ### The 6 parts, in short 1. Give ChatGPT your real numbers: Start a message, Paste the facts, Send it. 2. Check the math: Spot the gap, Ask for a fix, Send the correction. 3. Try other call volumes: Ask a what-if, Slow month, busy month, Send it. 4. Compare the setup effort: Bring in the setup, Describe what you built, Send it. 5. Build the launch checklist: Plan the launch, Ask for the checklist, Send it. 6. Write the greeting: One more thing, Ask for the greeting, Send it, Copy the greeting. **Start here first:** [Your First Voice Agent in ElevenLabs: An After-Hours Receptionist](https://learn.ynteractive.com/content/elevenlabs-voice-agent-tutorial-after-hours-receptionist) (module 1 of AI Voice Agents) **In the course:** [Part 11 · AI Voice Agents](https://learn.ynteractive.com/content/course/part-11-ai-voice-agents) (Module 54 of the [full course](https://learn.ynteractive.com/content/course)) ## The example Cedar Creek Plumbing is a made-up 6-person family plumbing company. It gets about 120 [after-hours calls](https://learn.ynteractive.com/content/glossary/after-hours-answering-service) a month, around 3 minutes each: 360 minutes. Its receptionist, Riley, triages the problem, gives a safety tip for a leak, books non-emergency visits through an [n8n](https://learn.ynteractive.com/content/tool/n8n) workflow, and takes messages. Earlier modules in this course build Riley in ElevenLabs Agents (modules 1 to 4) and in Vapi (modules 5 and 6). ## The prices used (checked September 29, 2026) - ElevenLabs Agents: Starter $6 a month with 75 minutes, Creator $22 a month with 275 minutes, Pro $99 a month with 1,238 minutes; extra minutes $0.08 each. The [AI model](https://learn.ynteractive.com/content/glossary/ai-model) (LLM) is billed separately at the model's cost; this module uses $0.01 a minute as a clearly labelled placeholder, to be replaced with your real usage. ElevenLabs charges no telephony fee; the phone provider bills you. - Twilio phone line for ElevenLabs: about $1.15 a month for a US local number plus about $0.0085 a minute for inbound calls. - Vapi: $0.05 a minute platform fee, plus the transcriber, model and voice at the providers' rates; Vapi's own estimate for the Balanced setup with the Clara voice is about $0.04 a minute. The first Vapi phone number is free (US, inbound only) and Vapi lists its own telephony as free. ## The math ElevenLabs at 360 minutes: the Creator plan is $22 plus 85 extra minutes at $0.08 ($6.80), so $28.80. The Starter plan plus extra minutes also comes to exactly $28.80, because the $16 difference between the plans buys exactly 200 minutes at $0.08; above 275 minutes a month the plan choice doesn't change the price. Add the LLM placeholder ($3.60) and the Twilio line ($1.15 + 360 × $0.0085 = $4.21): $36.61 a month, about 31 cents a call. Vapi at 360 minutes: 360 × ($0.05 + $0.04) = $32.40, with $0 for the free Vapi number: about 27 cents a call. A slow month of 40 calls (120 minutes) is about $12.97 on ElevenLabs (Starter) and $10.80 on Vapi; a busy month of 300 calls (900 minutes) is about $89.80 and $81.00. At every volume the difference is a few cents a call. So for a small business the price is close to a tie, and the decision comes down to setup, who runs it, and how you like to be billed. ## Why you give ChatGPT the numbers, and check its math The prompt in this module lists every price and tells ChatGPT to use only those numbers and show its math. Don't ask an AI chatbot what a service costs: prices change often and it can be out of date or simply wrong. Even with the numbers in hand, check its work. In the lesson, ChatGPT's first table leaves out the Twilio phone line for ElevenLabs, which makes the two platforms look identical. One short correction fixes it. ## Setup effort compared - Getting started: ElevenLabs puts the prompt, voice and model on one screen. Vapi has you pick a transcriber, model and voice, or start from a preset such as Balanced. - Knowledge: ElevenLabs has a built-in knowledge base for a price list or FAQ. Vapi can do it with files and a Query tool, with more setup. - Phone number: ElevenLabs needs a Twilio (or SIP) number that you import with the account SID and auth token. Vapi gives you a free US number in one click. - Actions: both call an n8n workflow to book visits (ElevenLabs [webhook](https://learn.ynteractive.com/content/lovable-n8n-web-app-ai-backend) tools, Vapi API Request tools), and both can transfer a caller to a person. - Billing: ElevenLabs is a monthly plan plus extra minutes; Vapi is per minute only, with $5 of free credit to start. ## Which one to pick For a small shop where the office manager will run it, ElevenLabs is usually the simpler start. Vapi suits you if you'd rather skip a Twilio account, pay only for the minutes you use, or swap the voice or model later. Either works. Don't run both on your real number. ## The launch checklist Whichever you pick, work through this before the first real caller: - Tell callers it's AI: the greeting says it's an AI assistant and that the call may be recorded, and it answers honestly if asked. - Recording and consent: rules vary by state, and some require every party's consent. Check yours with a lawyer or your trade association; this module is not legal advice. Decide how long recordings and transcripts are kept. - Never collect card numbers or other sensitive details on the call. - Fallbacks: real emergencies transfer to the on-call phone; a gas smell means leave the house and call the gas company or 911; a failed booking tool ends in a callback promise, never a fake confirmation; if the platform is down, your carrier forwards unanswered calls to a person or voicemail. - After-hours rules: forward the main number to the AI number only after hours, on weekends and holidays; keep holidays and the on-call rota in the prompt; only promise callback times the on-call plumber can keep; no prices or arrival times. - Test with 10 real phone calls (a leak, a gas smell, a booking, a price question, an angry caller, a wrong number), make sure a transfer really rings, and check bookings arrive. - In the first month, read a few call logs a week, compare the real cost per call with your estimate, and set spending alerts. ## Try it yourself, free Reading the steps is a start. Doing them is how it sticks. The interactive module walks you through every click in a practice copy of ChatGPT, checks each step, and never touches your real accounts. [Start “ElevenLabs vs Vapi: Cost per Call and a Launch Checklist”](https://learn.ynteractive.com/modules/voice-agents-compare-launch) ## Step by step: ElevenLabs vs Vapi: Cost per Call and a Launch Checklist All 19 steps of the interactive module, in order. In the module you watch each one, then do it yourself in a practice copy that checks your work. ### 1. Give ChatGPT your real numbers 1. **Start a message.** You have Riley working in both ElevenLabs and Vapi. Now decide which one to launch. Click the message box. You've built Riley, Cedar Creek Plumbing's after-hours receptionist, in both ElevenLabs and Vapi. Now it's time to pick one for real callers. ChatGPT is handy for the math, as long as you give it the numbers. 2. **Paste the facts.** Paste in the prompt. It gives the call volume and the prices checked on each pricing page, and tells ChatGPT to use only those numbers and show its math. It has Cedar Creek's call volume, about 120 after-hours calls a month at three minutes each, and the prices we checked on each pricing page. It tells ChatGPT to use only these numbers, never to guess a price, and to show its math. ```text I run Cedar Creek Plumbing, a small family plumbing company. We built the same after-hours AI phone receptionist on two platforms and need to pick one. Use ONLY the numbers below, show your math, and don't look up or guess any other prices. Our calls: about 120 after-hours calls a month, 3 minutes each on average. ElevenLabs Agents (prices checked Sept 29, 2026): - Starter $6/month includes 75 minutes; Creator $22/month includes 275 minutes - Extra minutes $0.08 each - The AI model (LLM) is billed separately: use $0.01 a minute as our placeholder - Phone line from Twilio: $1.15 a month plus $0.0085 a minute Vapi (prices checked Sept 29, 2026): - $0.05 a minute platform fee - Transcriber + model + voice: about $0.04 a minute (Vapi's estimate for our setup) - Free Vapi phone number, $0 a minute Give me a table: monthly cost and cost per call for each, then one sentence on what it means. ``` 3. **Send it.** Click the **send** arrow. ### 2. Check the math 4. **Spot the gap.** Both come out at $32.40. But look at the **Phone line** row for ElevenLabs: it is empty. Click the message box. Both come out at exactly 32 dollars 40 a month. Too neat? Look at the phone line row: ChatGPT left the Twilio line out for ElevenLabs. Always check an AI's math against your own numbers. 5. **Ask for a fix.** Paste in the correction. 6. **Send the correction.** Click the **send** arrow. Send it. ### 3. Try other call volumes 7. **Ask a what-if.** Now ElevenLabs is $36.61 a month (31 cents a call) and Vapi $32.40 (27 cents). Click the message box. Now it's right: about 31 cents a call on ElevenLabs and 27 cents on Vapi. But calls go up and down with the seasons. 8. **Slow month, busy month.** Paste in the question: the same math for 40 calls and for 300 calls a month. Ask for a slow month with 40 calls and a busy one with 300. 9. **Send it.** Click the **send** arrow. Send it. ### 4. Compare the setup effort 10. **Bring in the setup.** At every volume the gap is a few cents a call. Click the message box. At every volume, the difference is only a few cents a call. So price won't decide this. The setup and who runs it will. 11. **Describe what you built.** Paste in the notes. They list what each platform took in modules 1 to 6, and who will run it. They describe what each setup really took in this course: the Twilio account and import for ElevenLabs, the free number and per-minute billing in Vapi, the tools, and who will run it day to day. ```text Here's what setting each one up took us, from our own build: ElevenLabs: agent, prompt and voice in one screen; price list uploaded to the knowledge base; a Twilio account to buy a number, then import it with the SID and auth token; webhook tools to n8n; a built-in transfer-to-number tool. Vapi: pick a transcriber, model and voice (or a preset); a free Vapi number in one click, no Twilio account; API Request tools to n8n; versions for the assistant and the tools; pay per minute, no plan. Compare the setup effort in a short table, then recommend one for a 6-person shop where the office manager will run it. ``` 12. **Send it.** Click the **send** arrow. Send it. ### 5. Build the launch checklist 13. **Plan the launch.** Read the table and the recommendation. The choice is yours; either works. Click the message box. ChatGPT suggests ElevenLabs for a small shop, and Vapi if you'd rather skip Twilio or pay only per minute. It's your call, and either one works. What matters more is launching safely. 14. **Ask for the checklist.** Paste in the request: a launch checklist with five headings, and no legal advice. It asks for a checklist under five headings: telling callers it's AI and recording consent, fallbacks, after-hours rules, testing, and the first month. ```text Write our launch checklist for the AI receptionist, as short checkboxes under these headings: 1. Telling callers it's AI, and recording consent 2. Fallbacks (emergencies, tool failures, the platform being down) 3. After-hours rules (when it answers, holidays, what counts as an emergency) 4. Testing before launch 5. The first month after launch Don't give legal advice; tell us what to check and with whom. ``` 15. **Send it.** Click the **send** arrow. Send it. ### 6. Write the greeting 16. **One more thing.** Check the fallbacks: emergencies go to a person, a failed tool never gives a fake confirmation, and calls still reach someone if the platform is down. Click the message box. Read the fallbacks closely. Real emergencies go to a person, a failed booking never gets a fake confirmation, and calls still reach someone if the platform is down. And recording rules vary by state, so check yours with a lawyer or your trade association. Now the greeting. 17. **Ask for the greeting.** Paste in the request. Ask for the first thing Riley says: short, and it must say it's an AI and that the call may be recorded. ```text Last one: write the first thing the receptionist says, under 30 words. It must say it is an AI assistant for Cedar Creek Plumbing and that the call may be recorded. ``` 18. **Send it.** Click the **send** arrow. Send it. 19. **Copy the greeting.** Click the **Copy** icon under the answer, then paste the greeting into the First Message of whichever platform you launch. Copy the greeting, and paste it into the first message of whichever platform you pick. Then work through the checklist before the first real caller. ## Frequently asked questions **Is ElevenLabs or Vapi cheaper for a small business?** With prices checked on September 29, 2026, they are close. For 120 three-minute calls a month the example works out to about $36.61 on ElevenLabs (including a Twilio line and a placeholder for the AI model) and $32.40 on Vapi with a free Vapi number: 31 versus 27 cents a call. **Does ElevenLabs charge for phone calls?** ElevenLabs Agents includes telephony with no ElevenLabs fee, but your phone provider (for example Twilio) charges for the number and the minutes, and the AI model is billed separately. **Do I have to tell callers they are talking to an AI?** Say it anyway: it builds trust and avoids surprises. Whether you must, and how call recording consent works, depends on where you and your callers are, so check with a lawyer. This course is not legal advice. **Can I ask ChatGPT for the latest prices?** Don't rely on it. Copy the prices from each pricing page yourself, give them to ChatGPT, ask it to use only those numbers and show its math, and check the result. **Is this module free?** Yes. Stepthrough is completely free and the practice copy needs no accounts. [Start this module free](https://learn.ynteractive.com/modules/voice-agents-compare-launch) - Previous in AI Voice Agents: [Give Your Voice Agent Knowledge: A Price List and FAQ in ElevenLabs](https://learn.ynteractive.com/content/elevenlabs-agent-knowledge-base-price-list-faq) ## More in AI Voice Agents - [Your First Voice Agent in ElevenLabs: An After-Hours Receptionist](https://learn.ynteractive.com/content/elevenlabs-voice-agent-tutorial-after-hours-receptionist): Build an after-hours phone receptionist for a small plumbing company: system prompt, a greeting that says it is AI, a calm voice and a fast model, then test-call it in… [All 3 modules in AI Voice Agents](https://learn.ynteractive.com/content/course/ai-voice-agents) ## More ChatGPT tutorials - [Prompt Writing 101: Role, Context, Task, Format](https://learn.ynteractive.com/content/prompt-writing-101-role-context-task-format): Ask ChatGPT for the same customer email twice, once vaguely and once with a four-part prompt, and see exactly why structure gets a better answer. - [Get Your Business Recommended by ChatGPT](https://learn.ynteractive.com/content/get-recommended-by-chatgpt): Ask ChatGPT who it recommends in your town, see which pages it trusts, then use it to write an FAQ page and Business Profile copy that answer real customer questions. - [Summarize a Meeting Transcript with ChatGPT](https://learn.ynteractive.com/content/chatgpt-summarize-meeting-transcript): Paste a long meeting transcript into ChatGPT and get the decisions, an action items table with owners and due dates, and a follow-up email ready to send. - [Reply Faster: Turn a Messy Customer Email Thread into a Clear, Friendly Reply](https://learn.ynteractive.com/content/chatgpt-reply-to-customer-emails): Turn a messy customer email thread into a clear, friendly reply: give ChatGPT the thread and your facts, check every promise, fix the tone, and send it from Gmail. - [ChatGPT Projects: Keep Your Brand Voice, Files and Instructions in One Place](https://learn.ynteractive.com/content/chatgpt-projects-brand-voice): Set up a ChatGPT Project with your brand voice instructions, your price list and your style guide, so every new chat already knows your business. - [Analyze a Spreadsheet with ChatGPT: Upload a CSV, Get a Chart and Three Takeaways](https://learn.ynteractive.com/content/chatgpt-analyze-spreadsheet-csv): Upload a CSV of sales to ChatGPT, have it check the data first, ask your questions, get a chart you can download, and finish with the three takeaways that matter. ## Terms in this lesson Plain-English definitions, each with an example from a free lesson. [AI receptionist](https://learn.ynteractive.com/content/glossary/ai-receptionist) · [After-hours answering service](https://learn.ynteractive.com/content/glossary/after-hours-answering-service) · [Workflow](https://learn.ynteractive.com/content/glossary/workflow) · [AI model](https://learn.ynteractive.com/content/glossary/ai-model) · [Large language model (LLM)](https://learn.ynteractive.com/content/glossary/large-language-model) · [Prompt](https://learn.ynteractive.com/content/glossary/prompt) [The full AI glossary](https://learn.ynteractive.com/content/glossary) ## Browse related tutorials [AI Voice Agents (3)](https://learn.ynteractive.com/content/course/ai-voice-agents) · [ChatGPT tutorials (30)](https://learn.ynteractive.com/content/tool/chatgpt)