improving media generation docstrings

app.py

@@ -679,7 +679,7 @@ kokoro_interface = gr.Interface(
     api_description=(
         "Synthesize speech from text using Kokoro-82M. Returns (sample_rate, waveform) suitable for playback. "
         "Parameters: text (str), speed (float 0.5–2.0), voice (str). "
-        "Return the generated image to the user."
+        "Return the generated media to the user in this format `![Alt text](URL)`"
     ),
     allow_flagging="never",
 )
@@ -805,7 +805,7 @@ image_generation_interface = gr.Interface(
         "Generate a single image from a text prompt using a Hugging Face model (serverless Inference). "
         "Parameters: prompt (str), model_id (str, creator/model-name), negative_prompt (str), steps (int, 1–100), cfg_scale (float, 1–20), "
         "sampler (str, label only), seed (int, -1=random), width/height (int, 64–1216). Returns a PIL.Image. "
-        "Return the generated image to the user in this format `![Alt text](URL)`"
+        "Return the generated media to the user in this format `![Alt text](URL)`"
     ),
     allow_flagging="never",
 )
@@ -982,7 +982,7 @@ video_generation_interface = gr.Interface(
     api_description=(
         "Generate a short video from a text prompt using a Hugging Face model (Serverless Inference). "
         "Parameters: prompt (str), model_id (str), negative_prompt (str), steps (int), cfg_scale (float), seed (int), "
-        "width/height (int), fps (int), duration (float). Returns a file path to an MP4 that MCP exposes as a file URL."
+        "width/height (int), fps (int), duration (float). Return the generated media to the user in this format `![Alt text](URL)`"
     ),
     allow_flagging="never",
 )