Skip the LLM compose step for logged-out visitors

compose_song falls back to the built-in demo lyrics + structured caption
when no OAuth token and no HF_TOKEN secret are available, so anonymous
users can still run the generation pipeline.

README.md

@@ -25,7 +25,9 @@ wav) → Output Song / Seed Used / Stats → `make_video` (CPU ffmpeg visualizer
 Seed, randomize-seed, flow-matching steps, guidance scale, duration, and video title are editable reference
 nodes. `hf_oauth: true` lets the owner edit the canvas; visitors sign in to run the pipeline, and the
 `compose_song` LLM step uses the visitor's own OAuth token (`inference-api` scope), so inference bills the
-visitor directly. Set an `HF_TOKEN` secret if you want anonymous/local runs to fall back to the owner's token. The original Blocks app
+visitor directly. For logged-out visitors the LLM step is skipped (it's an add-on) and the pipeline
+generates from the built-in demo lyrics + structured caption; set an `HF_TOKEN` secret to let anonymous
+runs use the owner's token instead. The original Blocks app
 (with the live-streaming PCM player) is preserved as `app_blocks.py` — workflow fn nodes are plain callables,
 so chunk-by-chunk streaming is replaced by a final audio subject.
 

app.py

@@ -505,10 +505,7 @@ def _llm_json(system, user, token=None):
     # directly); fall back to the Space owner's HF_TOKEN secret for local/anonymous runs.
     api_key = getattr(token, "token", None) or os.environ.get("HF_TOKEN")
     if not api_key:
-        raise gr.Error(
-            "Sign in with Hugging Face to compose (the LLM step bills your own account), "
-            "or ask the Space owner to add an HF_TOKEN secret."
-        )
+        return None  # logged-out visitor, no owner secret: caller skips the LLM step
     client = OpenAI(base_url="https://router.huggingface.co/v1", api_key=api_key)
     completion = client.chat.completions.create(
         model="MiniMaxAI/MiniMax-M3:fastest",
@@ -526,6 +523,10 @@ def compose_song(description, duration, token: gr.OAuthToken = None):
     if not description.strip():
         raise gr.Error("Describe the song you want first.")
     data = _llm_json(_COMPOSER_SYSTEM, f"Song description: {description}\nTarget duration: {int(duration)} seconds.", token)
+    if data is None:
+        # No OAuth token (logged out) and no HF_TOKEN secret: the LLM compose step is an
+        # add-on — skip it and generate from the built-in demo lyrics + structured caption.
+        return DEFAULT_LYRICS, DEFAULT_GLOBAL, DEFAULT_VOCALS, DEFAULT_ARRANGEMENT
     return data["lyrics"], data["global_metadata"], data["vocal_details"], data["arrangement"]