What's New?

1. Responsive UI Polish: A round of UI adjustments to make the app work better across viewport sizes.

2. New Composer-Based Onboarding Flow: A new onboarding experience built on top of the Assistant Builder and powered by Composer is rolling out to select users as part of a phased release.

1. New Assistant Builder Experience: An updated, streamlined assistant configuration experience, now available to all users.
   • UI optimizations for the Phone Numbers page were also made to align its look and interactions with the new experience.

1. Soniox — General Availability: The Soniox transcriber is now rolled out to all customers. Configure it on any assistant via assistant.transcriber (provider: soniox) for low-latency, multilingual real-time speech-to-text.

1. Deepgram Flux — Multilingual Support: Full support for Deepgram’s multilingual Flux model. Multilingual agents can now leverage the same smart turn-taking that powers the English Flux transcriber, making cross-lingual conversations feel more fluid and natural.

1. Logs UX Refresh: New filter layout plus a round of UX improvements — improved date picker, active row is clearly highlighted across all log views when the flyout is opened, log tables are fully keyboard-accessible, sortable cost and duration columns, pagination, and more.

2. Squads contextEngineeringPlan Handoff Type — previousAssistantMessages: Forwards only the conversation history from before the current assistant’s session. The current assistant’s own messages and tool calls are excluded entirely from the handoff payload. See the updated handoff context configuration docs.

3. assistant.speechStarted Event — Live Captions & Word-Level Timing (GA): A new opt-in message fires as the assistant begins speaking each segment, carrying the full turn text, turn, source (model / force-say / custom-voice), and optional timing:

   • Per-word alignment on ElevenLabs
   • Cursor-based word-progress on Minimax (set voice.subtitleType: "word", with correct CJK handling)
   • Text-only fallback on all other providers

   Subscribe by adding "assistant.speechStarted" to your assistant’s clientMessages and/or serverMessages — now GA with no feature flag. Use it for live captions, karaoke-style highlighting, or any UI that needs to stay in sync with assistant audio. Fully backward-compatible; no existing messages changed.

4. Autofallbacks on Transcribers: Let Vapi pick the best transcriber to fall back to if your primary one fails — even mid-call. Opt in by setting assistant.transcriber.fallbackPlan.autoFallback.enabled to true. See the updated transcriber fallback plan docs.

1. Monitoring — GA: Automated call quality monitoring is now generally available. Detect issues with trigger-based rules, get alerts when something goes wrong, and surface resolution suggestions — all from the dashboard.

   • Monitoring quickstart
   • Announcement blog post

Here’s a summary of major items shipped from October 2025 through March 2026.

Platform

1. Squads v2: Visual builder to simplify sophisticated multi-assistant orchestration with seamless handoffs between specialized agents.

2. Composer (Alpha): Intelligent assistant inside the dashboard that allows you to describe what you need through plain text prompts to help build, adjust, and debug voice agents.

3. Simulations (Alpha): Voice agent testing feature to build confidence through enabling systematic, AI-powered testing in specific scenarios with evaluation of outcomes.

4. Monitoring & Issues: Automated call quality monitoring with trigger-based issue detection, alerting, and resolution suggestions.

5. HIPAA with Data Retention: New compliance mode with private storage and in-dashboard toggle/purchase flow — available for additional cost.

6. Zero Data Retention: Compliance mode that keeps context data during call as needed to execute tasks and retains no data afterwards.

7. Consolidated Logs: Unified log viewing into a single page.

8. Vapi Voices: 12 new ultra-realistic voices released, optimized for latency and cost with adjustable speed controls exposed. 8 legacy voices deprecated.

New Models & Provider Support

Transcriber Models (Speech-to-Text)

1. Deepgram Nova-3 Languages: Added Hebrew, Urdu, Tagalog, and Arabic bilingual support.

2. Cartesia Transcriber: ink-whisper.

3. Soniox: stt-rt-v4.

Intelligence Models (LLM)

1. GPT-5 Family: OpenAI’s latest intelligence models, including GPT-5, 5-Mini, 5-Nano, 5.1, 5.2, 5.4, 5.4-Mini, 5.4-Nano.

2. Claude 4.5–4.6: Anthropic’s latest intelligence models Sonnet 4.5, Opus 4.5, Opus 4.6, Sonnet 4.6.

3. Gemini 3 Flash: Google’s latest intelligence models.

4. Grok 4 Fast: Reasoning and non-reasoning variants.

5. GPT Realtime Mini: OpenAI’s lightweight realtime model.

Voice Models (Text-to-Speech)

1. Cartesia: sonic-3, sonic-3-2026-01-12, sonic-3-2025-10-27.

2. WellSaid: Caruso (new), legacy.

3. Inworld: inworld-tts-1 (REST, original), inworld-tts-1.5-max (WebSocket, $10/M chars), inworld-tts-1.5-mini (WebSocket, $5/M chars).

4. ElevenLabs Scribe v2: Latest version of ElevenLabs speech-to-text.

Developer Tools & API

1. Structured Outputs Improvements: Updates to our AI-powered analysis and data extraction tool, including transient structured outputs, audio-based extraction, and regex extraction.

2. SIP Request Tool + DTMF over SIP INFO: Send SIP requests and DTMF tones via SIP INFO messages during calls.

3. Variable Passing Between Tool Calls: Pass output variables from one tool call as input to subsequent tool calls.

4. Encrypted Tool Arguments: Encrypt sensitive tool arguments to protect data in transit.

5. Low Confidence Speech Hook: Hook that triggers when the transcriber returns low-confidence speech results.

6. Time Elapsed Hook: Hook that triggers at specified time intervals during a call.

7. assistant.speechStarted Event: New event fired when the assistant begins speaking.

8. MCP Improvements: Bearer auth, $ref dereferencing, child tool messages/discovery.

9. Warm Transfer Improvements: SIP support, caller ID, context engineering, variable filling.

1. Legacy Endpoint Removal: The following deprecated endpoints have been removed as part of our API modernization effort:

   • /logs - Use call artifacts and monitoring instead
   • /workflow/{id} - Access workflows through the main workflow endpoints
   • /test-suite and related paths - Replaced by the new evaluation system
   • /knowledge-base and related paths - Integrated into model configurations

2. Knowledge Base Architecture Change: The knowledgeBaseId property has been removed from all model configurations. This affects:

   • XaiModel, GroqModel, GoogleModel
   • OpenAIModel, AnthropicModel, CustomLLMModel
   • All other model provider configurations

3. Transcriber Property Deprecation: AssemblyAITranscriber.wordFinalizationMaxWaitTime and FallbackAssemblyAITranscriber.wordFinalizationMaxWaitTime are now deprecated:

   • Use smart endpointing plans for better speech timing control
   • More precise conversation flow management
   • Enhanced end-of-turn detection capabilities

4. Schema Path Cleanup: Removed numerous unused schema paths from model configurations to simplify the API structure and improve performance. This cleanup affects internal schema references but doesn’t impact your existing integrations.

5. New v2 API: We are introducing a new API version v2. These changes are part of our ongoing effort to:

   • Simplify the API structure for better developer experience
   • Remove redundant and deprecated functionality
   • Complete the transition to new evaluation and compliance systems
   • Improve API performance and maintainability

For details on the new features that replace these deprecated endpoints, see our recent changelog entries:

• Enhanced Authentication & Custom Credentials (Aug 30)
• Recording Consent & Compliance Management (Sep 2)
• Evaluation System Foundation (Sep 5)
• Evaluation Execution & Results Processing (Sep 28)

Migration Guide

Removed Endpoints

The following endpoints are no longer available:

• GET /logs - Use call artifacts instead
• GET /workflow/{id} - Use main workflow endpoints
• GET /test-suite, POST /test-suite - Use evaluation endpoints
• GET /test-suite/{id}, PUT /test-suite/{id}, DELETE /test-suite/{id} - Use evaluation management
• POST /test-suite/{testSuiteId}/run - Use evaluation runs
• GET /knowledge-base, POST /knowledge-base - Integrated into model configurations
• All related nested endpoints and operations

See Also:

• Authentication System Updates (Aug 30) - For credential management migration
• Recording Consent Features (Sep 2) - For compliance system details
• Enhanced Transcription (Sep 8) - For AssemblyAI timing alternatives

1. Evaluation Execution Engine: Run comprehensive assistant evaluations with EvalRun and CreateEvalRunDTO. Execute your mock conversations against live assistants and squads to validate performance and behavior in controlled environments.

2. Multiple Evaluation Models: Choose from various AI models for LLM-as-a-judge evaluation:

   • EvalOpenAIModel: GPT models including GPT-4.1, o1-mini, o3, and regional variants
   • EvalAnthropicModel: Claude models with optional thinking features for complex evaluations
   • EvalGoogleModel: Gemini models from 1.0 Pro to 2.5 Pro for diverse evaluation needs
   • EvalGroqModel: High-speed inference models including Llama and custom options
   • EvalCustomModel: Your own evaluation models with custom endpoints

3. Evaluation Results: Comprehensive result tracking with EvalRunResult:

   • status: Pass/fail evaluation outcomes
   • messages: Complete conversation transcript from the evaluation
   • startedAt and endedAt: Precise timing information for performance analysis

4. Target Flexibility: Run evaluations against different targets:

   • EvalRunTargetAssistant: Test individual assistants with optional overrides
   • EvalRunTargetSquad: Evaluate entire squad performance and coordination

5. Evaluation Status Tracking: Monitor evaluation progress with detailed status information:

   • running: Evaluation in progress
   • ended: Evaluation completed
   • queued: Evaluation waiting to start
   • Detailed endedReason including success, error, timeout, and cancellation states

6. Judge Configuration: Optimize evaluation accuracy with model-specific settings:

   • maxTokens: Recommended 50-10000 tokens (1 token for simple pass/fail responses)
   • temperature: 0-0.3 recommended for LLM-as-a-judge to reduce hallucinations

Evaluation Capabilities

1. Enhanced Beep Detection: Improve voicemail detection accuracy with CreateVoicemailToolDTO.beepDetectionEnabled specifically for Twilio-based calls. This feature detects the characteristic beep sound that indicates voicemail recording has started.

2. Workflow Voicemail Integration: Configure comprehensive voicemail handling in workflows with enhanced message and detection capabilities:

   • Workflow.voicemailMessage: Custom messages for voicemail scenarios (up to 1000 characters)
   • Workflow.voicemailDetection: Configurable detection methods for different providers

3. Assistant Voicemail Enhancement: Improved voicemail handling in assistant configurations with Assistant.voicemailMessage and Assistant.voicemailDetection for consistent behavior across all conversation types.

4. Multiple Detection Methods: Choose from various voicemail detection providers:

   • Google: GoogleVoicemailDetectionPlan for AI-powered detection
   • OpenAI: OpenAIVoicemailDetectionPlan for intelligent voicemail recognition
   • Twilio: TwilioVoicemailDetectionPlan for carrier-level detection
   • Vapi: VapiVoicemailDetectionPlan for integrated detection

5. Beep Detection for Call Flows: The new beep detection capability works specifically with Twilio transport, providing reliable voicemail identification when traditional detection methods may not be sufficient.

6. Voicemail Tool Configuration: Enhanced tool rejection and messaging capabilities ensure appropriate handling when voicemail is detected, with configurable responses based on your business requirements.

Voicemail Management Features