feat: Text generation suggestions via Gemini models

Backend:
- New text_gen.py module with Gemini text generation client
  - suggest_prompts(): generates 3 prompt suggestions from user input
  - get_text_gen_config(): loads config from .env (model, API key, safety)
  - get_available_models(): returns list of Gemini text models
  - Safety settings: harassment, hate_speech, sexually_explicit, dangerous
    Configurable via TEXT_GEN_SAFETY_* env vars (default: BLOCK_ONLY_HIGH)
- New endpoints:
  * POST /api/text-gen/suggest — generates 3 prompt suggestions
  * GET /api/text-gen/models — lists available text models
  * GET /api/text-gen/config — returns safe config (API key masked)
  * POST /api/text-gen/config — updates config in-memory
- Default model: gemini-2.5-flash
- Fallback API key: LYRIA_API_KEY if TEXT_GEN_API_KEY not set
- System prompt: expert music producer persona for rich musical detail

Frontend:
- New SuggestionsModal component for displaying generated suggestions
  - Auto-fetches when opened
  - Shows loading spinner, error state, or suggestion list
  - Each suggestion clickable to populate the prompt field
  - Hover effect: 'Click to use this prompt →'
- Generator component updated with 'Get Suggestions ✨' button
  - Appears next to prompt label in both Basic and Advanced modes
  - Disabled when no input present
  - Clicking suggestion populates the prompt/structured fields
- SettingsModal expanded with Text Generation section (collapsible)
  - Same pattern as Image Generation: model, API key status, safety filters
  - Editable dropdowns with Save Changes button
  - Refactored to use shared useGenSettings hook for both sections

Models Available:
- gemini-2.5-flash (GA, default) — best price-to-performance
- gemini-2.5-pro (GA) — highest quality reasoning
- gemini-3.5-flash (Preview) — next-gen Flash
- gemini-3.1-flash (Preview) — latest with thinking

Architecture KB and Tasks KB updated.

backend/src/main.py

@@ -14,9 +14,8 @@ from database import init_db, create_clip, get_clips, get_clip, delete_clip, upd
 from lyria_client import generate_clip, save_clip_file
 from waveform import generate_waveform_svg, get_waveform_path
 from analysis import analyze_clip
-from image_gen import generate_artwork, get_image_gen_config, get_available_models
-from analysis import analyze_clip
 from image_gen import generate_artwork, get_image_gen_config, get_available_models, SAFETY_LEVELS
+from text_gen import suggest_prompts, get_text_gen_config, get_available_models as get_text_models
 
 app = FastAPI(title="Lyria Studio", version="0.1.0")
 
@@ -320,6 +319,55 @@ async def update_image_gen_configuration(request: dict):
     
     return {"status": "updated"}
 
+# ─── Text Generation Endpoints ──────────────────────────────────────
+
+@app.get("/api/text-gen/models")
+async def list_text_gen_models():
+    """List available text generation models."""
+    return {"models": get_text_models()}
+
+@app.get("/api/text-gen/config")
+async def get_text_gen_configuration():
+    """Get current text generation configuration (safe values only)."""
+    config = get_text_gen_config()
+    return {
+        "model": config["model"],
+        "safety_harassment": config["safety_harassment"],
+        "safety_hate_speech": config["safety_hate_speech"],
+        "safety_sexually_explicit": config["safety_sexually_explicit"],
+        "safety_dangerous": config["safety_dangerous"],
+        "has_api_key": bool(config.get("api_key")),
+    }
+
+@app.post("/api/text-gen/config")
+async def update_text_gen_configuration(request: dict):
+    """Update text generation configuration (in-memory only)."""
+    if "model" in request:
+        os.environ["TEXT_GEN_MODEL"] = request["model"]
+    if "safety_harassment" in request:
+        os.environ["TEXT_GEN_SAFETY_HARASSMENT"] = request["safety_harassment"]
+    if "safety_hate_speech" in request:
+        os.environ["TEXT_GEN_SAFETY_HATE_SPEECH"] = request["safety_hate_speech"]
+    if "safety_sexually_explicit" in request:
+        os.environ["TEXT_GEN_SAFETY_SEXUALLY_EXPLICIT"] = request["safety_sexually_explicit"]
+    if "safety_dangerous" in request:
+        os.environ["TEXT_GEN_SAFETY_DANGEROUS"] = request["safety_dangerous"]
+    return {"status": "updated"}
+
+@app.post("/api/text-gen/suggest")
+async def get_suggestions(request: dict):
+    """Get prompt suggestions based on user input."""
+    user_input = request.get("input", "").strip()
+    if not user_input:
+        raise HTTPException(status_code=400, detail="Input is required")
+    
+    try:
+        config = get_text_gen_config()
+        suggestions = await suggest_prompts(user_input, config=config)
+        return {"suggestions": suggestions}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Suggestion generation failed: {str(e)}")
+
 @app.patch("/api/clips/{clip_id}")
 async def rename_clip_endpoint(clip_id: str, request: RenameRequest):
     clip = await get_clip(clip_id)

backend/src/text_gen.py

@@ -0,0 +1,140 @@
+"""Text generation client using Google Gemini models for prompt suggestions."""
+
+import os
+from typing import Optional
+
+from google import genai
+from google.genai import types
+
+# ── Configuration ──────────────────────────────────────────────────────────
+
+def get_text_gen_config():
+    """Load text generation configuration from environment."""
+    return {
+        "api_key": os.environ.get("TEXT_GEN_API_KEY") or os.environ.get("LYRIA_API_KEY"),
+        "model": os.environ.get("TEXT_GEN_MODEL", "gemini-2.5-flash"),
+        "safety_harassment": os.environ.get("TEXT_GEN_SAFETY_HARASSMENT", "BLOCK_ONLY_HIGH"),
+        "safety_hate_speech": os.environ.get("TEXT_GEN_SAFETY_HATE_SPEECH", "BLOCK_ONLY_HIGH"),
+        "safety_sexually_explicit": os.environ.get("TEXT_GEN_SAFETY_SEXUALLY_EXPLICIT", "BLOCK_ONLY_HIGH"),
+        "safety_dangerous": os.environ.get("TEXT_GEN_SAFETY_DANGEROUS", "BLOCK_ONLY_HIGH"),
+    }
+
+# ── Safety Settings ────────────────────────────────────────────────────────
+
+def build_safety_settings(config: dict) -> list:
+    """Build safety settings from config."""
+    return [
+        types.SafetySetting(
+            category="HARM_CATEGORY_HARASSMENT",
+            threshold=config.get("safety_harassment", "BLOCK_ONLY_HIGH"),
+        ),
+        types.SafetySetting(
+            category="HARM_CATEGORY_HATE_SPEECH",
+            threshold=config.get("safety_hate_speech", "BLOCK_ONLY_HIGH"),
+        ),
+        types.SafetySetting(
+            category="HARM_CATEGORY_SEXUALLY_EXPLICIT",
+            threshold=config.get("safety_sexually_explicit", "BLOCK_ONLY_HIGH"),
+        ),
+        types.SafetySetting(
+            category="HARM_CATEGORY_DANGEROUS_CONTENT",
+            threshold=config.get("safety_dangerous", "BLOCK_ONLY_HIGH"),
+        ),
+    ]
+
+# ── Prompt Suggestion ──────────────────────────────────────────────────────
+
+SYSTEM_PROMPT = """You are an expert music producer and prompt engineer for AI music generation models.
+Your task is to help users write better prompts for generating music with AI models like Google Lyria.
+
+When given a user's idea or partial prompt, expand it into detailed, evocative music descriptions.
+Include specific details about:
+- Genre and subgenre
+- Instrumentation and sound design
+- Mood, energy, and emotional arc
+- Tempo and rhythmic feel
+- Structure (intro, build, drop, breakdown, etc.)
+- Production style (clean, lo-fi, distorted, etc.)
+
+Keep suggestions concise but rich in musical detail. Use terminology that music producers would use.
+"""
+
+async def suggest_prompts(user_input: str, model: str = None, config: dict = None) -> list:
+    """Generate prompt suggestions based on user input.
+
+    Returns a list of suggestion strings.
+    """
+    cfg = config or get_text_gen_config()
+    client = genai.Client(api_key=cfg["api_key"])
+    
+    safety = build_safety_settings(cfg)
+    
+    prompt = f"""The user wants to generate music with an AI music model. They described their idea as:
+
+"{user_input}"
+
+Please provide 3 distinct, detailed prompts they could use. Each prompt should be a single paragraph that vividly describes the music. Make them different from each other — explore different angles of the user's idea.
+
+Format your response as a numbered list (1., 2., 3.) with just the prompts, no extra commentary."""
+
+    response = client.models.generate_content(
+        model=model or cfg["model"],
+        contents=prompt,
+        config=types.GenerateContentConfig(
+            system_instruction=SYSTEM_PROMPT,
+            safety_settings=safety,
+            temperature=0.9,
+            max_output_tokens=2048,
+        ),
+    )
+
+    # Parse response into list of suggestions
+    text = response.text.strip()
+    suggestions = []
+    
+    # Split by numbered lines (1., 2., 3.)
+    import re
+    # Find all text after numbers
+    matches = re.findall(r'(?:^|\n)\d+\.\s*(.*?)(?=\n\d+\.|$)', text, re.DOTALL)
+    if matches:
+        suggestions = [m.strip() for m in matches if m.strip()]
+    else:
+        # Fallback: split by newlines and filter empty
+        lines = [line.strip() for line in text.split('\n') if line.strip()]
+        # Remove lines that are just numbers
+        suggestions = [line for line in lines if not re.match(r'^\d+\.$', line)]
+    
+    return suggestions[:3]  # Ensure max 3
+
+# ── Available Models ───────────────────────────────────────────────────────
+
+AVAILABLE_TEXT_MODELS = [
+    {
+        "id": "gemini-2.5-flash",
+        "name": "Gemini 2.5 Flash",
+        "description": "Best price-to-performance. Fast, capable, 1M context.",
+        "status": "GA",
+    },
+    {
+        "id": "gemini-2.5-pro",
+        "name": "Gemini 2.5 Pro",
+        "description": "Highest quality reasoning and coding. Slower, more expensive.",
+        "status": "GA",
+    },
+    {
+        "id": "gemini-3.5-flash",
+        "name": "Gemini 3.5 Flash",
+        "description": "Next-gen Flash. Strong reasoning, fast inference.",
+        "status": "Preview",
+    },
+    {
+        "id": "gemini-3.1-flash",
+        "name": "Gemini 3.1 Flash",
+        "description": "Latest Flash with thinking capabilities.",
+        "status": "Preview",
+    },
+]
+
+def get_available_models() -> list:
+    """Return list of available text generation models."""
+    return AVAILABLE_TEXT_MODELS