Releases: GetStream/Vision-Agents
Releases · GetStream/Vision-Agents
v0.1.13 - Add support for Moondream Detection
What's Changed
- Moondream Detection API by @Nash0x7E2 in #136
- feat: add AWS Bedrock function calling implementation by @d3xvn in #120
- Extract model download logic to utils by @Nash0x7E2 in #146
V0.1.12 - Deepgram Flux, vogent, smart turn
- Deepgram flux (with v2 listen)
- Vogent turn detection
- Smart turn detection
- Logging cleanup
New Contributors
- @filiplajszczak made their first contribution in #116
Full Changelog: v0.1.11...v0.1.12
Support Sonic-3 by default
- Add support for
sonic-3by default in Cartesia TTS
OpenRouter support v0.1.9
Fish Audio support
Fish audio support
Add AWS support - V0.1.7
First version of AWS bedrock and AWS nova sonic support.
Bug fix - Chat, Deepgram version update
- Change the default chat channel type to
messaging - Bump Deepgram plugin to version 5
v0.1.0 - First release
The first release of vision agents. With a focus on vision AI, low latency and completely open. Quickly build video agents with any model or video edge network.
- Processors: for providing state and modifying/publishing video and audio
- Turn keeping
- TTS
- STT
- LLM
- Realtime LLM (gemini and openAI)
First release. First step of many towards 1.0.