feat(realtime): Initial Realtime API implementation

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

core/config/backend_config.go

@@ -37,6 +37,7 @@ type BackendConfig struct {
 	TemplateConfig      TemplateConfig         `yaml:"template"`
 	KnownUsecaseStrings []string               `yaml:"known_usecases"`
 	KnownUsecases       *BackendConfigUsecases `yaml:"-"`
+	Pipeline            Pipeline               `yaml:"pipeline"`
 
 	PromptStrings, InputStrings                []string               `yaml:"-"`
 	InputToken                                 [][]int                `yaml:"-"`
@@ -72,6 +73,18 @@ type BackendConfig struct {
 	Options []string `yaml:"options"`
 }
 
+// Pipeline defines other models to use for audio-to-audio
+type Pipeline struct {
+	TTS           string `yaml:"tts"`
+	LLM           string `yaml:"llm"`
+	Transcription string `yaml:"transcription"`
+	VAD           string `yaml:"vad"`
+}
+
+func (p Pipeline) IsNotConfigured() bool {
+	return p.LLM == "" || p.TTS == "" || p.Transcription == ""
+}
+
 type File struct {
 	Filename string         `yaml:"filename" json:"filename"`
 	SHA256   string         `yaml:"sha256" json:"sha256"`