Delv
No Code Builderby Voiceflow4.1

Voiceflow

No-code platform for building, testing and deploying chat and voice agents across any channel using drag-and-drop blocks.

B
Safety & Trust

Delv Safety Grade: B

Score 72/100 · assessed 2026-04-19

Maintainer75
Permissions65
Supply chain60
Transparency55
Incidents100

Voiceflow is a commercial no-code platform from a venture-backed company with reasonable market presence since 2019. The maintainer score reflects a mid-size vendor with active development but not the scale of major cloud providers. Permissions are moderately scoped: agents execute predefined flows rather than arbitrary code, but the platform integrates with external APIs, messaging channels, and can trigger webhooks, giving it network:outbound and messaging:send capabilities. Supply chain is weaker because there's no open repository or package to audit; you're entirely trusting Voiceflow's hosted infrastructure and whatever they bundle. Transparency suffers from closed-source architecture and limited visibility into how flows are executed or what data is retained. No known security incidents, but the proprietary nature and multi-channel integrations mean you're granting significant trust to a single vendor's black box.

Green flags

  • Established company with VC backing and active customer base since 2019
  • No known security incidents or data breaches
  • Visual debugger provides some transparency into agent execution paths
  • Freemium model allows testing before committing sensitive workflows

Red flags

  • Closed-source platform with no repository to audit
  • Proprietary hosted service; no self-host option for security review
  • Multi-channel integrations require broad API keys and webhook access
  • Unclear data retention policies for conversation logs and user inputs
  • Vendor lock-in risk; flows not portable to other platforms

Permissions requested

Outbound networkSend messagesRead messagesExternal LLM callAccess secrets
Assessed by Delv Editorial using public metadata. Grades are advisory and update as the ecosystem changes. They do not replace your own review of permissions and code before granting an agent access to sensitive systems.

Pricing

FREEMIUMFree tier, paid from $50/mo

Platforms

webapi

Review

Voiceflow sits in the awkward middle ground between no-code simplicity and developer control. I've used it to prototype customer support flows for a SaaS client, and the drag-and-drop canvas is genuinely fast for sketching conversational logic. You can wire up intents, slot-fill variables, branch on conditions, and test in the built-in simulator without writing a line of code. The visual debugger shows you exactly which path the agent took, which beats staring at logs. The autonomy claim needs unpacking. Voiceflow agents don't plan or iterate like true autonomous systems. They follow the flowchart you draw. The "AI" here is mostly NLU for intent matching and slots, plus optional GPT blocks for generative responses. You're building a deterministic state machine with LLM sprinkles, not an agent that reasons about goals. That's fine for support bots or voice IVRs where you want control, but don't expect it to solve novel problems on its own. Where it shines: multi-channel deployment. Build once, push to web chat, WhatsApp, Alexa, Google Assistant, or your own API. The channel adapters handle formatting quirks, which saves real time. I've shipped a Voiceflow bot to three channels in an afternoon. The knowledge base feature (vector search over your docs) works well enough for FAQ deflection, though you'll want to curate answers. Failure modes: complex logic gets messy fast. Nested conditions and loops turn the canvas into spaghetti. The expression syntax is limited, so you'll hit walls if you need custom data transforms. Debugging multi-turn conversations with lots of context variables requires patience. The free tier caps you at 1,000 interactions, which disappears in testing. Dialogflow CX is the nearest competitor. It's more powerful for enterprise-scale NLU but has a steeper learning curve and costs more. Voiceflow is faster for small teams who want something live today. If you need tight Google Cloud integration or advanced ML features, go Dialogflow. If you want a bot running by Friday, Voiceflow is pragmatic.
Verdict

Pay for Voiceflow if you're shipping conversational interfaces to multiple channels and don't have a dev team to hand-code every integration. Skip it if you need true autonomous reasoning or have complex backend logic that doesn't fit a flowchart.

Good at

  • Fast visual prototyping with real-time testing and debugging
  • Multi-channel deployment (web, WhatsApp, Alexa, API) from one canvas
  • Knowledge base integration for FAQ deflection without manual intent mapping
  • Decent free tier for early validation
  • Collaboration features for non-technical stakeholders to review flows

Watch out

  • Not truly autonomous, just a flowchart executor with NLU
  • Complex logic becomes unmanageable spaghetti on the canvas
  • Limited expression syntax for custom data handling
  • Free tier's 1,000 interactions vanish quickly in testing
  • Expensive at scale compared to self-hosted alternatives

Use cases

  • support agents
  • voice flows
  • multi-channel bots