Voiceflow Pricing Plans and the Cost Per Message, Phone Minute and Model

Voiceflow pricing has two parts: a plan fee that third-party reviews report at $60 a month for Pro and $150 for Business (about 10 percent less billed yearly), and usage that Voiceflow publishes in its own credits table at $0.005 per user message, $0.05 per phone minute, and model tokens at 1.25 times the provider list price. The plan fee comes back as the same dollar amount of usage credit, so what you really pay is usage. On our assumptions a four message support chat costs about $0.03 on GPT-5 mini, $0.05 on Claude Haiku 4.5 and $0.11 on Claude Sonnet 4.6, and a phone minute lands around $0.11 to $0.17 all in.

How much does Voiceflow cost per month?

Start with what changed. The public pricing page at voiceflow.com/pricing, read in September 2026, no longer prints a plan price. It offers two doors: "For Agencies & Partners", with a free trial and "transparent, usage-based billing", and "For Businesses", with "Book a demo" and "Request pricing". The self-serve plan prices now live inside the product on the Plans and Billing screen, and the figures below are the ones third-party reviews recorded from that screen in September 2026.

The seat line is the one small teams underestimate. A four person team on Pro pays $60 for the plan plus $150 for three extra editors, so $210 a month before a single customer message, and only $60 of that comes back as usage credit. If only one person builds and the rest review, keep them as free viewers.

What does Voiceflow charge per message, minute and model?

This part is first-party. Voiceflow's documentation publishes a credits pricing table, and every figure here comes from it. The platform itself charges two things: $0.005 every time a user sends a message or a request hits the Conversations API, and $0.05 per minute of phone call. Model calls are billed on top, per 1,000 tokens.

The pattern holds across every model we checked: Voiceflow bills tokens at exactly 25 percent over the provider's list rate, and several OpenAI models carry a higher "priority" rate on top of that. We could not find this multiplier stated in words anywhere; it only shows up when you divide the two tables. Voice adds its own lines from the same table, for example Deepgram speech to text at $0.005 a minute and ElevenLabs text to speech at $0.045 per 1,000 characters.

How much does a Voiceflow chat conversation cost?

Our assumptions: a support chat has four user messages, and each reply sends about 4,000 input tokens (instructions, retrieved knowledge and history) and 400 output tokens. Change the model and the bill moves more than any plan decision you will make.

Put next to the no-code chat tools, this is cheap. The same Sonnet class answer costs about $0.10 to $0.11 per reply through Chatbase message credits, and a Tidio Lyro conversation is $0.70 or more, as our Tidio alternatives comparison shows. Botpress, the closest builder platform, now charges $0.50 to $0.65 per conversation with $0.10 of AI usage included, as our Botpress pricing breakdown works through. Voiceflow is closer to raw API cost than either, which is why it tends to attract teams that are building something serious rather than dropping a widget on a site.

How much does a Voiceflow phone call cost per minute?

Voiceflow's own charge is the $0.05 platform minute. Speech recognition, voice and model tokens come from the same credits table, and the phone line itself does not: Voiceflow connects to your own Twilio or Vonage number, which third-party reviews put at $0.01 to $0.03 a minute. Assuming three agent turns a minute and about 450 spoken characters:

• On Claude Haiku 4.5: $0.05 platform, $0.005 Deepgram, about $0.02 ElevenLabs, about $0.0225 model, so roughly $0.10 in Voiceflow credit, or $0.11 to $0.13 with the phone line.
• On Claude Sonnet 4.6: the model share rises to about $0.0675, so roughly $0.14 in credit, or $0.15 to $0.17 all in.

That sits in the same band as the voice specialists. Vapi also charges $0.05 a minute of hosting and lands at about $0.09 to $0.14 all in, and Retell AI at about $0.13 to $0.25, per our Retell vs Vapi pricing breakdown. If phone is the main channel, compare all three on concurrency and latency rather than on the per-minute line; the full field is on our Vapi alternatives page.

What happens when Voiceflow credits run out?

Several third-party reviews report that on self-serve plans the agent simply stops answering when the month's credit is used up, and that the fix is moving up a plan rather than buying a top-up. We could not confirm the current overage rule on a public Voiceflow page, so check the Plans and Billing screen before launch, and set a usage alert at 70 to 80 percent of the credit. An agent that goes silent on the busiest day of the month costs more than any plan upgrade.

Is Voiceflow worth it, or should you build your own?

On cost alone, Voiceflow is hard to beat at small and medium volume. Its margin over raw APIs is the $0.005 message fee plus the 25 percent token markup, which on our Sonnet chat works out to about $0.038 per conversation ($0.11 against about $0.072 direct). You would need to build and host the equivalent yourself to save that, and the build still needs the same upkeep a Voiceflow agent does. Spread a $3,600 to $24,000 build over 24 months and add $50 to $150 of hosting, and the markup only pays for the move at about 5,000 to 30,000 conversations a month. On Haiku, where the markup is about $0.026 a chat, the range moves out to roughly 7,700 to 44,000.

So for most teams the real decision is not Voiceflow or custom. It is who builds the Voiceflow agent, which model it runs on, and how many editor seats you pay for. A well built agent that answers on Haiku and escalates to Sonnet only when it has to can cut the usage line in half. Voice agents also make calls to your own APIs in the middle of a conversation, and when one of those endpoints is down the caller hears silence, so uptime monitoring on the API endpoints your agent calls belongs in the launch checklist.

If you want that build done by a specialist, our chatbot developers page covers what to screen for and what they charge, and for phone agents the AI voice agent development page covers the extra work telephony adds. Before any agent goes live, run it through the checks in how to evaluate an AI agent before you ship it; Voiceflow's own evaluation tools help, but they grade against the tests you write.

Voiceflow pricing questions buyers ask

Is Voiceflow free?

Voiceflow offers a free trial with no credit card for the agency and partner route. Production use runs on a paid plan, reported at $60 a month for Pro and $150 for Business, with usage billed from the credits table. Businesses that want a managed deployment are routed to "Request pricing".

Does Voiceflow charge per message?

Yes. Voiceflow's credits table charges $0.005 every time a user sends a message or a request reaches the Conversations API, excluding launch events. Model tokens, speech and voice are charged on top, so the message fee is usually the smaller part of the cost of a reply on anything larger than a mini model.

Can I bring my own model to Voiceflow?

Voiceflow lists OpenAI, Anthropic, Google, Groq and xAI models in its credits table, plus its own models, and its product pages mention bringing your own model. Check your plan for the terms; through the built-in integrations, tokens are billed at the 1.25x rate shown above.

How much does Voiceflow Enterprise cost?

Voiceflow does not publish it. Vendr's buyer data, as reported by third-party reviews, puts the median contract at about $258,625 a year on a $34,500 platform fee that includes three editor licenses. Treat that as an order of magnitude, not a quote, and get usage rates in writing, since usage is most of the bill.