fix header for pcm

tools/server/README.md

@@ -806,7 +806,7 @@ Returns raw audio bytes (`audio/wav` by default) rather than JSON. For more info
 
 `n_predict`: Max number of audio frames to generate. Defaults to `512`; generation normally stops earlier once the model emits an end-of-speech token.
 
-`response_format`: `wav` (default) or `pcm` (raw `float32` samples, no header).
+`response_format`: `wav` (default) or `pcm` (raw `float32` little-endian mono samples, no header; the response `Content-Type` carries the sample rate, e.g. `audio/pcm;rate=24000;encoding=float;bits=32`).
 
 `stream`: If `true`, the response is streamed as audio becomes available instead of waiting for the full generation to finish. WAV streaming writes an RFC-noncompliant header with an unknown (`0xFFFFFFFF`) size field, since the final length isn't known up front; most players and decoders (ffmpeg, VLC, ...) handle this by reading until EOF.
 

tools/server/server-context.cpp

@@ -5216,6 +5216,8 @@ void server_routes::init_routes() {
             return res;
         }
 
+        const auto info = mtmd_gen_audio_get_info(ctx_server.mctx);
+
         const json body = json::parse(req.body);
 
         std::string prompt = json_value(body, "input", json_value(body, "prompt", std::string()));
@@ -5283,7 +5285,10 @@ void server_routes::init_routes() {
         task.id = rd.get_new_id();
         rd.post_task(std::move(task));
 
-        const std::string content_type = response_format == "pcm" ? "audio/L16" : "audio/wav";
+        // raw float32 LE mono samples; audio/L16 would imply 16-bit big-endian (RFC 2586)
+        const std::string content_type = response_format == "pcm"
+            ? "audio/pcm;rate=" + std::to_string(info.sample_rate) + ";encoding=float;bits=32"
+            : "audio/wav";
 
         if (!stream) {
             auto result = rd.next(req.should_stop);