Back to Vocalis AI
Product Changelog

What's New in Vocalis AI

Continuous updates to our realtime voice engine, SIP mesh, and developer CLI.

v2.4.0

Semantic End-of-Turn v3 & Native Salesforce Streaming

Major reduction in premature cut-offs during multi-part customer account numbers and real-time bi-directional CRM sync.

  • FeatureReleased Semantic End-of-Turn v3 model reducing false interruptions by 41% on long numeric sequences.
  • FeatureAdded zero-config OAuth connector for Salesforce Service Cloud live screen-pops on warm SIP transfers.
  • ImprovementReduced Frankfurt and Singapore edge PoP synthesis time-to-first-byte by 38ms.
v2.3.2

Real-Time PCI DTMF Masking & CLI Replay Debugger

Enhanced security controls for payment intake workflows and interactive waveform inspection inside the terminal CLI.

  • FeatureAdded inline RFC 2833 DTMF tone suppression for PCI-DSS compliant payment capture.
  • ImprovementUpgraded @vocalis-ai/cli with split-channel latency breakdown histograms.
  • FixResolved rare Opus packet-loss concealment artifact during 4G cellular handoffs.
v2.3.0

Multilingual Code-Switching & Custom Pronunciation Lexicons

Seamless mid-call language switching between English, Spanish, French, and German without re-initializing WebRTC sessions.

  • FeatureIntroduced dynamic code-switching ASR + TTS routing within a single active call stream.
  • ImprovementAdded SSML-free phoneme lexicon editor for medical and fintech brand terminology.
  • FixFixed webhook retry backoff jitter when downstream ERP endpoints return HTTP 429.

Upgrade to the latest Vocalis CLI

npx @vocalis-ai/cli init --template=enterprise-voice

Deploy v2.4