Chatbase vs Voiceflow, compared on price, phone and set-up

Chatbase sells message credits on a public price list. A reply costs 1 to 6 credits, depending on the model you pick, and an action can double that by making a second request. Credits renew on the 1st of each calendar month, and whatever’s left is lost. The free plan has 50 credits a month. Hobby, at $40 a month, or $32 a month billed yearly, has 700, which is between 116 and 700 replies, or fewer when actions run. Voice and phone calls come with Standard, at $150 a month, and spend the same credits by the minute. Chatbase pricing goes through what a credit buys on each plan.

Voiceflow’s pricing page names no plan and no price. Agencies get a free trial with no card, and businesses are asked to book a demo. Its billing documentation lists four plans, Free, Pro, Business and Enterprise, and says to open Plans and Billing inside Voiceflow to see what they cost. Free is a one-time grant of credits that doesn’t renew. Paid plans get a monthly allowance that resets each billing cycle, and extra editor seats and phone numbers are add-ons.

What Voiceflow does publish is the rate card for its credits. Each message a visitor sends costs $0.005, so 1,000 messages spend $5 before the model has written a word. The model’s tokens are charged by the thousand on top, at a rate for each model, and a phone call costs $0.05 a minute plus its speech-to-text and text-to-speech. Even refreshing your knowledge base can spend credits, if the source uses an AI chunking strategy. So you can price a conversation on Voiceflow, but not a month, until you’ve signed up.

Choose Chatbase if you want an agent on your site this week, at a price you can read before you sign up. Its guide puts building an agent and getting the embed code at about 5 minutes, and the same agent reaches 16 channels, from WhatsApp to Slack. When you outgrow Hobby, Standard adds the helpdesk, voice, phone calls and the API for $150 a month.

Choose Voiceflow if you’re designing an agent that must follow set steps, such as verifying a customer before it discusses their account, or if the phone is your main channel. It makes and takes calls, forwards a call to a person when it needs one, and gives you environments, A/B tests and AI scoring of your transcripts. It also suits agencies, with white-labelling and client handoff tools. You’ll need to sign up or book a demo to see the price.

Both are built for someone whose job is the agent, tuning its model, instructions and channels. If you’re a founder who answers support between everything else, there’s a third option that asks much less of you.

You probably don’t need a conversation designer. You need the same few questions answered without you, so you can build. Deacon answers them in seconds, day or night, from your own docs and in whatever language your customer writes in. If your docs don’t cover a question, it says it doesn’t know instead of guessing, and asks for the visitor’s email. When they leave one or ask for a person, you get an email, and you reply from Deacon’s dashboard. There are no playbooks, workflows or versions to publish, just one script tag.

It gets better as you answer. Every question your docs didn’t cover sits in the dashboard with an Answer this button, and what you write becomes part of what Deacon knows. When you save it, Deacon asks itself the question again, shows you its new reply and tells you how many other open questions your answer covers. Over time, it needs you less.

And it shows you what your customers are stuck on and what they want next. Every conversation is saved and searchable by what visitors asked. Topics collect similar questions, count the people who asked and label most of them, as a bug report, a feature ask, a pricing question and so on. The same script also runs website analytics, with your visitors, their sources, goals and funnels, so there’s no second script to add.

You can see the whole price before you start. The plans are on the pricing page, every feature is on every plan, and they differ only in answers, knowledge, pages from your website and seats. Every plan uses the same model, so there are no per-message rates to add up. Nothing is billed past your plan’s answers, and no plan upgrades itself. The free plan renews every month, needs no card and has no end date, and no plan carries a badge to pay to remove.