v1.6.1 v0.13.0
supermemoryai/supermemorySep 2, 2026 → Sep 4, 20262 days between releases
What's New
Major divergence: CLI tools replaced by SIP/WebRTC infrastructure
Breaking Changes
- Complete functional overhaul replacing CLI tools with SIP/WebRTC infrastructure
New Features
- Configurable SIP proxy for outbound requests
- Speechmatics STT integration
- New REST and VSI endpoints for managing custom leg data
- SIPREC metadata validation
- WS ingress playout lead
Improvements
- Configurable SIP proxy for outbound requests
- Improved audio frame handling in mixer
- Fixed TLS URI parsing to prevent crashes
- Fixed RTP port recycling to prevent leaks
- Fixed Contact URI on outgoing register
Bug Fixes
- Fixed MCP call errors no longer quarantining the connection
- Fixed TLS URI parsing
- Fixed RTP port recycling
- Fixed Contact URI on outgoing register
- Stop double-pacing the leg-agent audio path
Upgrade Notes
These releases belong to different projects (Strix and VoiceBlender), so there is no direct upgrade path between them.
The Two Releases
v1.6.1
Sep 2, 2026· 2 months ago
This release improves stability and configuration management, including merging environment variables into CLI config and fixing environment checks on the TUI start screen. It updates the recommended list of models and fixes MCP connection handling.
Full release notesv0.13.0
Sep 4, 2026· 2 months ago
This release enhances SIP and WebRTC capabilities with a configurable proxy, improved audio frame handling, and the integration of Speechmatics for speech recognition. It also introduces new APIs for managing custom leg data and fixes critical issues with TLS parsing and RTP port recycling.
Full release notes