diff --git a/core/http/endpoints/localai/tts.go b/core/http/endpoints/localai/tts.go
index ca3f58bd9e2..7c73c6337a0 100644
--- a/core/http/endpoints/localai/tts.go
+++ b/core/http/endpoints/localai/tts.go
@@ -9,16 +9,19 @@ import (
 	"github.com/gofiber/fiber/v2"
 	"github.com/mudler/LocalAI/core/schema"
 	"github.com/rs/zerolog/log"
+
+	"github.com/mudler/LocalAI/pkg/utils"
 )
 
 // TTSEndpoint is the OpenAI Speech API endpoint https://platform.openai.com/docs/api-reference/audio/createSpeech
-//	@Summary	Generates audio from the input text.
-//  @Accept json
-//  @Produce audio/x-wav
-//	@Param		request	body		schema.TTSRequest	true	"query params"
-//	@Success	200		{string}	binary				"generated audio/wav file"
-//	@Router		/v1/audio/speech [post]
-//	@Router		/tts [post]
+//
+//		@Summary	Generates audio from the input text.
+//	 	@Accept json
+//	 	@Produce audio/x-wav
+//		@Param		request	body		schema.TTSRequest	true	"query params"
+//		@Success	200		{string}	binary				"generated audio/wav file"
+//		@Router		/v1/audio/speech [post]
+//		@Router		/tts [post]
 func TTSEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, appConfig *config.ApplicationConfig) func(c *fiber.Ctx) error {
 	return func(c *fiber.Ctx) error {
 
@@ -67,6 +70,13 @@ func TTSEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, appConfi
 		if err != nil {
 			return err
 		}
+
+		// Convert generated file to target format
+		filePath, err = utils.AudioConvert(filePath, input.Format)
+		if err != nil {
+			return err
+		}
+
 		return c.Download(filePath)
 	}
 }
diff --git a/core/schema/localai.go b/core/schema/localai.go
index cdc3e5b0784..1b663ae0baa 100644
--- a/core/schema/localai.go
+++ b/core/schema/localai.go
@@ -32,6 +32,7 @@ type TTSRequest struct {
 	Voice    string `json:"voice" yaml:"voice"` // voice audio file or speaker id
 	Backend  string `json:"backend" yaml:"backend"`
 	Language string `json:"language,omitempty" yaml:"language,omitempty"` // (optional) language to use with TTS model
+	Format   string `json:"response_format,omitempty" yaml:"response_format,omitempty"` // (optional) output format
 }
 
 type StoresSet struct {
diff --git a/pkg/utils/ffmpeg.go b/pkg/utils/ffmpeg.go
index 16656d8e7c0..68683370823 100644
--- a/pkg/utils/ffmpeg.go
+++ b/pkg/utils/ffmpeg.go
@@ -4,6 +4,7 @@ import (
 	"fmt"
 	"os"
 	"os/exec"
+	"strings"
 )
 
 func ffmpegCommand(args []string) (string, error) {
@@ -23,3 +24,32 @@ func AudioToWav(src, dst string) error {
 	}
 	return nil
 }
+
+// AudioConvert converts generated wav file from tts to other output formats.
+// TODO: handle pcm to have 100% parity of supported format from OpenAI
+func AudioConvert(src string, format string) (string, error) {
+	extension := ""
+	// compute file extension from format, default to wav
+	switch format {
+	case "opus":
+		extension = ".ogg"
+	case "mp3", "aac", "flac":
+		extension = fmt.Sprintf(".%s", format)
+	default:
+		extension = ".wav"
+	}
+
+	// if .wav, do nothing
+	if extension == ".wav" {
+		return src, nil
+	}
+
+	// naive conversion based on default values and target extension of file
+	dst := strings.Replace(src, ".wav", extension, -1)
+	commandArgs := []string{"-y", "-i", src, "-vn", dst}
+	out, err := ffmpegCommand(commandArgs)
+	if err != nil {
+		return "", fmt.Errorf("error: %w out: %s", err, out)
+	}
+	return dst, nil
+}
diff --git a/swagger/docs.go b/swagger/docs.go
index c283dcb0052..284656bf134 100644
--- a/swagger/docs.go
+++ b/swagger/docs.go
@@ -1721,7 +1721,11 @@ const docTemplate = `{
                 "voice": {
                     "description": "voice audio file or speaker id",
                     "type": "string"
-                }
+                },
+		"response_format": {
+                    "description": "(optional) output format of generated audio file, defaults to wav, accept wav, mp3, flac, aac, opus",
+                    "type": "string"
+                },
             }
         },
         "schema.ToolCall": {
diff --git a/swagger/swagger.json b/swagger/swagger.json
index 0a3be179759..3ec82878525 100644
--- a/swagger/swagger.json
+++ b/swagger/swagger.json
@@ -1714,6 +1714,10 @@
                 "voice": {
                     "description": "voice audio file or speaker id",
                     "type": "string"
+                },
+		"response_format": {
+                    "description": "(optional) output format of generated audio file, defaults to wav, accept wav, mp3, flac, aac, opus",
+                    "type": "string"
                 }
             }
         },
@@ -1742,4 +1746,4 @@
             "in": "header"
         }
     }
-}
\ No newline at end of file
+}
diff --git a/swagger/swagger.yaml b/swagger/swagger.yaml
index 7b6619b4873..af6055a12ab 100644
--- a/swagger/swagger.yaml
+++ b/swagger/swagger.yaml
@@ -679,6 +679,9 @@ definitions:
       voice:
         description: voice audio file or speaker id
         type: string
+      response_format:
+        description: (optional) output format of generated audio file, defaults to wav, accept wav, mp3, flac, aac, opus
+        type: string
     type: object
   schema.ToolCall:
     properties: