10 Best Voiceflow Alternatives of 2026

Botpress is the most direct swap for a conversation designer who likes Voiceflow's way of working but wants a public price list. The Agent Studio canvas lets you lay out an agent visually and drop into code where the logic needs it, while knowledge bases and Botpress Tables give it documents and structured data to work with. Paid plans include unlimited AI agents, and every conversation carries $0.10 of AI usage covered by Botpress, so model costs are folded into the conversation price. Plus costs $150 a month billed annually for 250 conversations, with extra packs of 100 for $65. The catch for voice projects is plan placement: the pricing table lists the voice channel only on Enterprise, where three minutes of a call count as one conversation. Teams that mostly ship web chat and WhatsApp agents get the best fit here; phone-first teams should price Enterprise early or look at Retell AI.

Retell AI suits conversation designers whose Voiceflow work has moved toward the phone. Call flows are drawn on a drag-and-drop canvas, and each node can call functions, transfer a call, collect DTMF digits or send an SMS. Simulation testing lets you run an agent against scripted callers before it answers real ones, and post-call analytics and AI quality assurance score production calls afterward. Pricing is the clearest in the voice category: pay as you go runs $0.07 to $0.31 per minute depending on the model, voice and telephony you pick, chat agents start at $0.002 per message, and new accounts get $10 in credits with 20 concurrent calls. The trade-off is scope. Retell is a voice company first, so designers who need rich web chat widgets or many messaging channels will find less there than in Botpress or Cognigy. Enterprise adds SSO, custom retention and higher concurrency at custom prices.

Dialogflow CX gives conversation designers a state-machine model that maps well to the way Voiceflow projects are planned: flows hold pages, pages hold state handlers, and generative fallback catches what the script misses. Generative playbooks sit alongside for goal-driven turns, with their own tools, examples and evaluations. Voice is first class. The docs cover speech models, speech adaptation, DTMF for telephony and voice agent design practices, and calls arrive through the Dialogflow CX Phone Gateway, partners such as Avaya, Twilio and Voximplant, or a SIP trunk. Prebuilt components handle common collection steps like dates, addresses and card numbers. Pricing is pay as you go at $0.007 per chat request and $0.001 per second of voice audio for Flows, with Playbooks at $0.012 and $0.002, and new users get $600 and $1,000 in trial credits. Expect to involve a developer for webhooks and Google Cloud projects.

Copilot Studio is the natural move for a design team inside a Microsoft shop. You can build chat or voice agents in one studio and publish them to Teams, websites and social channels. Microsoft's documentation describes two kinds of voice agent: basic voice agents that use topics and classic orchestration with natural language understanding, and real-time voice agents that use a speech-to-speech model for low-latency, natural turns. Both accept speech and DTMF input and support call transfer, which covers most IVR work a Voiceflow designer would hand over. Release controls are a strength, with dev, test and prod environments, version control and approval gates, plus automated evaluations. Billing is the weak point for planning: capacity comes in tenant-wide packs of 25,000 Copilot Credits for $200 a month, or pay as you go, and each action burns a varying number of credits. A phone number has to come from Azure Communication Services.

Cognigy is where a Voiceflow design team lands when the program outgrows a startup tool and has to live inside an enterprise contact center. Agents are designed in AI Agent Studio, which mixes LLM reasoning with deterministic steps in the same conversation, and you can describe an agent in plain language to get a draft persona, skills and knowledge to edit. Knowledge AI grounds answers in uploaded documents, synced storage and scraped sites. Before launch, AI Agent Evaluation runs simulated conversations with personas, missions and success criteria, and the AI Ops Center watches the fleet in production. Cognigy lists 100+ supported languages, 25K+ concurrent interactions and 110+ prebuilt tools and integrations, with contact-center connections for Avaya, AWS, Genesys, NiCE, Microsoft and 8x8. The downside is access: there is no public price or self-serve plan, so evaluation runs through a demo and sales process.

Rasa is the alternative for teams that left Voiceflow because they need agents, logic and data to stay on their own infrastructure. Designers can work in a no-code flow builder while engineers extend the same assistant in code and a custom actions server. Responses can vary by language and channel through a built-in content management system, and testing covers end-to-end runs, a live test panel and a conversation view for analysis. Deployment targets Kubernetes through Helm, with Redis for concurrency, Kafka for conversation data and OpenTelemetry for observability. The free Developer Edition can run in production for one bot with up to 1,000 external conversations a month. Voice is possible but not bundled: Rasa sells an IVR connector to AudioCodes VoiceAI Connect as an add-on, and channel connectors otherwise cover REST and WebSocket. Expect real engineering effort compared with any hosted studio here.

Typebot is the budget pick for conversation designers whose Voiceflow projects were scripted chat flows rather than voice agents. Its canvas uses 20+ blocks for messages, inputs, conditions and integrations, and every plan allows custom JavaScript and CSS plus webhooks, so a flow can call your own APIs. Billing counts chats, not messages: one visitor who sends ten messages is one chat. The free plan includes unlimited typebots and 200 chats a month, Starter costs $39 a month for two seats and 2,000 chats, and Pro at $89 adds five seats, 10,000 chats, WhatsApp, custom domains and deeper analytics. The limits are clear too. Typebot publishes no voice channel, so phone agents need another tool, and its strengths are guided flows and lead forms rather than LLM agents with knowledge bases. Plans are monthly only because pricing follows chat usage.

Kore.ai suits enterprises that want conversation design to become a governed, repeatable process across many agents. Agent Studio offers visual and code-based authoring in one workspace, and Arch, the built-in AI architect, turns a plain-language description into agents, workflows, tools, policies and handoffs. Under the hood, agents are defined in ABL, a typed language that makes behavior and guardrails explicit, and Kore.ai says policies are enforced outside the LLM. Autoloop then evaluates agents against goals you set and repairs the step that caused a miss, using StateTrace execution traces as evidence. Channel connectors cover voice, web, Slack, Teams and mobile, and Kore.ai also sells an agentic contact center product. For a Voiceflow designer this is a big step up in rigor and a big step up in process: there is no public price or self-serve plan, and the site is mid-rebrand, so some pages are split between old and new names.

Amazon Lex is the AWS-native route for teams whose voice agents end up in Amazon Connect. A single bot handles speech and text, and you choose between request and response interactions, where each user input is a separate API call, and streaming conversations, where one call listens across many turns. Pricing is plain usage: $0.004 per speech request and $0.00075 per text request, and AWS's own example puts 8,000 speech and 2,000 text requests at $33.50 for a month. That makes it one of the cheapest ways to run a phone bot at modest volume. For conversation designers the trade-off is tooling. The console is built for configuring intents and slots rather than drawing flows, and fulfillment usually lives in AWS Lambda, so a developer is part of the team from day one. It is a practical choice when the company's contact center already runs on AWS.

Synthflow is a phone-first platform for companies that would rather have a vendor help them launch than build every flow themselves. Agents answer inside existing phone systems through built-in connections for Zoom Phone, Five9, RingCentral, 8x8, Twilio, Telnyx, Vonage and Freshcaller, or through Synthflow's own telephony, where each call shows its SIP ladder and quality scores. During a call the agent reads and writes CRM, help desk and calendar data through 200+ integrations, lets callers interrupt or switch language, and passes a full transcript on warm transfer. Releases go out step by step with approval and are tested on your own call cases first. Web chat is available in all regions, while SMS and WhatsApp agents are limited to US workspaces. For a Voiceflow designer used to self-serve, the model is very different: Synthflow now sells only annual enterprise agreements, with no free plan, trial or per-minute rate.