This May, we made a massive leap forward with Workflows, our most requested feature for building complex, visual conversation flows. We also shipped dozens of API improvements, new models, and launched two exciting community programs.
Here's everything we shipped this month:
🎯 Major Feature Launch
Vapi Workflows: The new Workflows now lets you visually build complex conversation flows using a node-based interface ,no more cramming everything into single prompts.
What you can now do:
- Visual Flow Builder: Design conversation paths with conditional logic, sentiment routing, and dynamic branching
- Per-Node Customization: Different voices, languages, and models for each conversation step
- Smart Routing: AI-driven conditions that route based on user input, API responses, or detected sentiment
- Global Jump Nodes: "Talk to human" escape hatches available at any conversation point
- Multi-Assistant Orchestration: Route between specialized assistants for different tasks
Perfect for call centers, appointment scheduling, patient onboarding, and any complex conversation flow that single-prompt assistants struggle with.
Explore Workflows Documentation →
🤖 New Models and Voices
We expanded your options with new AI models and more voice personalities:
New Models:
- Azure OpenAI GPT 4.1: Including gpt-4.1-2025-04-14, gpt-4.1-mini-2025-04-14, and gpt-4.1-nano-2025-04-14
- Groq Models: meta-llama/llama-4-maverick-17b, mistral-saba-24b, compound-beta variants
- Anthropic: claude-opus-4-20250514, claude-sonnet-4-20250514 with thinking capabilities
- Gemini: gemini-2.5-flash-preview-05-20, gemini-2.5-pro-preview-05-06
New Voices:
- Kylie Voice: Available in Vapi voice provider
- Sesame Voice: New voice provider option for enhanced personality (update: deprecated)
Check your Vapi Dashboard to explore these options.
🛠️ Developer Tools & API Enhancements
- Enhanced Server Messages: Flexible array input for serverMessages. Configure exactly the message types you need.
- Advanced Endpointing Control: New smartEndpointingPlan and customEndpointingRules give you precise control over when your assistant stops listening and starts responding.
- Chat & Session APIs: New /chat and /session endpoints enable programmatic chat management and session persistence across multiple interactions.
- GoHighLevel Integration: Native tools for contact management and calendar scheduling:
- Contact creation and retrieval
- Calendar event scheduling and availability checking
- Regional Model Support: Specify regions for OpenAI models (e.g., gpt-4.1-2025-04-14:westus) for data residency compliance.
🎉 Community Programs
Vapi Build Challenge (Concluded): Our first-ever build challenge wrapped up with incredible submissions from the community. Winners showcased innovative voice AI applications across healthcare, customer service, and creative use cases.
Vapi Creators Program (New): Join our new program for builders, educators, and content creators shaping the future of voice AI.
Get early access, priority support, and revenue opportunities based on your content reach.
Apply to Vapi Creators Program →
⚡ Platform Improvements
- Workflow Entry Points: Start calls directly with workflowId ,same flexibility as assistants and squads [Docs]
- Warm Transfer Mode: New warm-transfer-experimental with hold music and voicemail detection [Docs]
- Enhanced Cost Tracking: New KnowledgeBaseCost type for tracking knowledge base query expenses [Docs]
- API Request Tool: Create custom HTTP request tools for any external API integration
- Extended Timeouts: Server timeout limits increased to 300 seconds for complex operations
What are the next big hurdles you're facing in production or development? Where should we focus our efforts next?
Let us know in the Discord or here.
As always, we're eager to see what you build.
\
