Release v0.7.8

Release Date: Dec 24, 2022

New Features
  • performance improvements to feature-server and sbc-inbound
  • add custom header on Refer indicating whether sbc-inbound should fix up the Refer-To
  • include service_provider_sid in call webhook
  • add support for custom azure voices (TTS)
  • return transcript faster if we get an exact match to a provided hint on an interim transcript (requires env JAMBONZ_GATHER_EARLY_HINTS_MATCH=1)
  • enable running multiple instance of node.js apps on EC2 for increased scalability
  • add support for Azure audio logging in gather and transcribe
  • add support for fromHost in createCall rest api
  • add ability to specify user or host part of From uri on outdial
  • update node images to the latest and most secure
  • update google speech libs
  • add env JAMBONES_RECORD_ALL_CALLS to enable global call recording
  • write stats for db lookup
  • include application_sid in cdr
Bug fixes
  • db caching had side affects of using closed http requestors
  • when handing over from wss to http close the wss socket
  • #192: config with dtmf only followed later by gather with speech not working
  • continuous asr timer in gather should not start until transcript is received
  • error handling in amd
  • microsoft tts voice was not being sent in tts request, resulting in a default voice being selected
  • #186 unhandled error when amd webhook returns non-success status code
  • ws error max connections error causes a crash
  • conference verb will have '_' property when leg is moved from other FS
  • move call leg to a different FS using the special REFER, we now must include X-Account-Sid header
  • account returned when carrier default routing is used was wrong
  • #66 support for gateways sending TLS with SRTP but not dtls
  • allow users to fully specify a Refer-To uri without us overiding it
  • fix for dtmf handler when running multiple instances on same EC2
  • support carrier domain in invite
  • block media going back to Five9 voicestream
  • include X-Account-Sid header when moving call between FS
  • when releasing media we were restarting ICE which we dont want to do
  • incorrectly making additional attempts after caller hung up or other failures on first attempt
  • strip X-Preferred-From-User and X-Preferred-From-Host from outgoing call
  • invalid uri in register causes crash
  • revamp algorithm to determine whether to become the active regbot in a cluster
SQL changes

No schema changes were made in this release.


Questions? Contact us at