← Blog

Botpress vs Voiceflow: Support Platform or Voice-First Agent Builder?

November 8, 2026

AI Agents & ToolsTool Comparisons

Botpress vs Voiceflow comes down to one question: is your agent mainly a support bot that lives in chat, or does it need to answer the phone? Botpress now sells itself as an AI-native customer support platform, with its own helpdesk and public, conversation-based plans. Voiceflow sells itself as an agent platform for CX teams, with chat, voice and SMS from one builder and real phone plumbing (Voiceflow numbers, Twilio, Vonage, Telnyx and SIP).

The short version:

  • Pick Botpress if you want a chat-first support agent, a built-in helpdesk, and a price you can work out from the pricing page before you talk to anyone.
  • Pick Voiceflow if phone calls are on the roadmap, or your CX team and engineers need to share one canvas with dev, staging and production environments.
  • Pick neither if you need to host the bot on your own servers. Both are hosted platforms today.

Every product claim below was checked on botpress.com and voiceflow.com this week. Here's the side-by-side first.

Quick comparison

BotpressVoiceflow
How it describes itselfAI-native customer support platformChat and voice agent platform for CX teams
BuildersStudio (visual), Viber (natural language), ADK (TypeScript)Studio with playbooks and workflows, Atlas copilot, API, CLI and MCP
HelpdeskYes, Botpress Desk, or connect Zendesk, Intercom, HubSpot, SalesforceNo own helpdesk; hands off to live-agent platforms
Main channelsWebchat, WhatsApp, Slack and other integrationsChat widget, voice, SMS, API
Phone callsVoice channel listed under EnterpriseInbound and outbound calls, own numbers, Twilio, Vonage, Telnyx, SIP
Pricing modelPlans by monthly conversations, AI usage allowance includedPlan fee plus credits, with seat and phone-number add-ons
Public price listYesAgency offer is self-serve; business pricing on request
Self-hostingNo. v12 and all self-hosted editions are sunsetNo. Private cloud hosting on Enterprise

Positioning: a support platform vs an agent builder

Spend five minutes on each homepage and the difference is obvious.

Botpress has moved a long way from "open-source chatbot framework." Its homepage now leads with "The AI-native customer support platform" and pitches agents that take real action on refunds, account updates and multi-step workflows, with escalation and handoff to a human when needed. It also sells Botpress Desk, a full ticket system, so you can run the AI agent and the human queue from one vendor. If you already have a helpdesk, Botpress says it works on top of it instead.

Voiceflow never had a helpdesk and isn't trying to build one. It describes one suite for building, testing and shipping chat and voice agents, aimed at product, CX and engineering teams working together. Its strength is the agent layer: how the agent reasons, which tools it calls, and how you test and release changes.

What that means for you: if "the bot" and "the support desk" are the same project in your company, Botpress fits the shape. If the agent is one component in a larger contact center, Voiceflow fits better.

How you build: Studio, Viber and ADK vs playbooks and workflows

Botpress gives you three ways in

Botpress now has three building surfaces, all on the same plans:

  • Botpress Studio, the visual agent builder most long-time users know.
  • Botpress Viber, a conversational builder. In its Viber docs, you describe what you want to a builder agent called Vibe, and it turns that into agent configuration you can then review screen by screen. (Not to be confused with the Viber messaging app, which Botpress also integrates with.)
  • Botpress ADK, a TypeScript framework and CLI. You write conversations, workflows, tools and actions as code, run them locally, then deploy to Botpress Cloud.

That spread is useful. A support manager can start in Viber, and an engineer can take over in the ADK when the logic gets serious.

Voiceflow mixes goal-based AI with fixed steps

Voiceflow's model is playbooks plus workflows. Playbooks are goal-based: you tell the agent what outcome to reach, and it navigates the conversation. Workflows are deterministic steps with AI reasoning where you want it, which is what you need for identity checks, payments, or anything a compliance team will audit.

Around that sit Atlas, an AI copilot for building agents, plus tests, simulations and LLM-based evaluations. Engineers can skip the canvas entirely: Voiceflow says you can work in the app, from the API, CLI and MCP, and its homepage invites you to set it up from your own coding agent.

Where Voiceflow pulls ahead: release management. Its pricing page describes environments that run from development to staging to production, all hosted on Voiceflow. Phone numbers can even be pointed at a draft version for testing by phone before you publish.

Voice and phone: the biggest real difference

This is where most "Botpress vs Voiceflow" decisions actually get made.

Voiceflow is built for phone calls. Its docs on connecting a phone number cover:

  • Voiceflow-provisioned US or Canadian numbers (its billing docs say one Voiceflow number per workspace is included with your plan)
  • Bringing your own number from Twilio, Vonage or Telnyx
  • Inbound and outbound calls through your agent
  • Inbound and outbound SMS, with Twilio numbers named explicitly for texting
  • Optional call recording, with a reminder that consent laws are your responsibility

If your numbers already live in a contact center, Voiceflow also lets you connect your phone system over SIP, so the agent answers some calls and transfers others without moving your numbers. Each plan also has a concurrent-calls limit you can raise with an add-on.

Botpress is chat-first. Its pricing page lists a voice channel among the Enterprise plan's features, and its FAQ says three minutes of a voice call counts as one conversation. So voice exists, but it's an enterprise conversation, not something you switch on during a free trial.

The call: if a phone agent is part of the plan in the next year, start with Voiceflow. If you're fairly sure you'll stay on web chat and WhatsApp, this difference doesn't matter much, and Botpress's support features matter more.

Pricing models: conversations vs credits

Both vendors change prices often, so treat this as a description of how they bill, and check the live pages before you budget.

Botpress bills by conversation

The Botpress pricing page lists a free plan, Plus and Team plans that include a set number of conversations per month, and custom Enterprise pricing. A few details are easy to miss:

  • What counts: any exchange with at least two messages from the end user in a billing month. A conversation that spans two months counts in both.
  • AI usage is included. Paid plans carry an AI-usage allowance, so model costs aren't a second bill until you pass it. After that, Botpress buys small AI-spend credit grants for you automatically, sold at cost.
  • Overages are automatic on paid plans. Near your limit, Botpress adds packs of 100 conversations (the Free plan has no top-ups or overages). Its FAQ says you can't turn this auto-recharge off, only get warned at 75% and 100%.
  • Seats: Free and Plus include a small number of seats. Team removes the cap.
  • Same price for every builder. Studio, Viber, Desk and ADK all run on the same plans.

Botpress also argues on that page that conversation pricing beats per-resolution pricing because your bill doesn't rise as the bot gets better. That's a fair point if you're comparing against helpdesk-native agents that charge per resolution.

Voiceflow bills by plan plus credits

Voiceflow's public pricing page splits into two offers. Agencies and partners get a free trial and usage-based billing with optional auto top-ups. Businesses book a demo and request pricing.

Its billing docs fill in the rest. A Voiceflow bill is a monthly plan fee, optional add-ons (extra editor seats, extra phone numbers, concurrent calls), and credit-based usage. Credits are spent when your agent does things like generating AI responses and making phone calls, and Voiceflow publishes a credits pricing table that lists every charged action. The docs name four plans (Free, Pro, Business and Enterprise), and the Free plan's credits are a one-time grant that doesn't renew, so it's for testing only. Viewers are free and unlimited.

Billing questionBotpressVoiceflow
Unit you pay forConversationsCredits (AI responses, phone calls and other actions)
Model costAllowance included in planConsumes credits
Running overAutomatic packs, can't be disabledOptional auto top-ups
Extra peopleSeats capped on lower plansPaid editor seats, free viewers
VoiceEnterprise; 3 minutes = 1 conversationCalls consume credits, plus phone-number and concurrency add-ons

What this means in practice: Botpress is easier to forecast, because you mostly need one number, your monthly conversation count. Voiceflow gives you more knobs, including your choice of model, but you'll need to estimate credits per conversation before you can forecast.

Hosting, control and lock-in

Neither tool is a self-hosted option anymore.

Botpress's docs say Botpress v12 has been officially sunset and is no longer available for purchase, download or new deployments, and the same goes for all other self-hosted or locally installed versions. Existing v12 subscribers are still supported. New Botpress projects run on Botpress Cloud.

Voiceflow is hosted too. Its billing docs list private cloud hosting as an Enterprise feature, along with single sign-on and custom credit allotments. That's the closest either vendor gets to running in your own environment.

On model lock-in, Voiceflow is more explicit: it says you can choose among major LLM providers or bring your own model. Botpress folds model costs into an included AI-usage allowance, which is simpler to budget.

If data truly can't leave your servers, look at frameworks you host yourself. Our Botpress alternatives guide covers Dify, Rasa and n8n for exactly that case.

Which one should you choose?

Run through these in order. The first one that clearly applies usually decides it.

  1. You need phone calls or SIP into an existing contact center. Voiceflow.
  2. You want the AI agent and the ticket queue from one vendor. Botpress, with Botpress Desk.
  3. You want to see the price before talking to sales. Botpress. Voiceflow's business pricing is on request.
  4. CX staff and engineers will edit the same agent, with staged releases. Voiceflow.
  5. One engineer, one support lead, chat only. Either works. Botpress is faster to price and start; Voiceflow gives you more room if voice shows up later.

If you're still not sure either is right, step back a level. Our Voiceflow alternatives guide covers phone-first tools like Retell and Vapi, and our customer service chatbots roundup covers helpdesk-native agents that skip the builder altogether.

Where Taku fits

Botpress and Voiceflow are for building a customer-facing agent you'll operate for years. Plenty of AI work isn't that: you just want to run a useful app or workflow someone else already built, on your own files. Taku is an AI-native desktop workspace for that, where you mirror a working setup, remix it and keep it. It won't replace a production support bot. Taku is in Beta, and the Mac app is available now.

FAQ

Is Botpress better than Voiceflow?

For chat-first customer support with a built-in helpdesk and public pricing, Botpress is usually the better fit. For voice agents, phone numbers, SIP or staged releases across a large team, Voiceflow is. Neither is better across the board.

Does Botpress support voice?

Yes, but its pricing page lists the voice channel under the Enterprise plan, and three minutes of a call counts as one conversation. Lower plans focus on webchat, WhatsApp and other messaging channels.

Does Voiceflow work with Twilio?

Yes. Voiceflow's docs let you connect a Twilio number with your Account SID and Auth Token for calls and SMS. It also supports Vonage, Telnyx, its own US and Canadian numbers, and SIP.

Can you self-host Botpress or Voiceflow?

Not for new projects. Botpress sunset v12 and all self-hosted editions, though existing v12 subscribers stay supported. Voiceflow is hosted, with private cloud hosting available on its Enterprise plan.

How is Botpress pricing different from Voiceflow pricing?

Botpress charges by monthly conversations and includes an AI-usage allowance. Voiceflow charges a plan fee plus credits for usage such as AI responses and phone calls, with add-ons for seats and phone numbers. Botpress publishes its plans; Voiceflow's business pricing is on request.

Which is easier for non-developers?

Both have visual builders. Botpress Viber lets you describe the agent in plain language, and Voiceflow has its Atlas copilot. Voiceflow's playbooks-plus-workflows model takes a little longer to learn but gives more control over scripted steps.