refactor: Phase 10 — Ollama service, image cache, config, frontend orphans

Final cleanup phase of the MCP-first pivot.

docker-compose:
  - docker-compose.yml: drop ollama service + OLLAMA_URL/MODEL env vars +
    IMAGE_CACHE / VAPID env comments
  - docker-compose.prod.yml: drop ollama service + Ollama env + GPU
    reservation
  - docker-compose.quickstart.yml: drop ollama service + Ollama env +
    GPU-reservation comment; quickstart instructions now point at the
    MCP Access tab instead of model-pull

Config:
  - Drop OLLAMA_URL, OLLAMA_MODEL, OLLAMA_BACKGROUND_MODEL,
    OLLAMA_KEEP_ALIVE_*, OLLAMA_NUM_CTX, EMBEDDING_MODEL (fastembed
    is hard-coded inside services/embeddings.py)
  - Drop IMAGE_CACHE_DIR, IMAGE_MAX_BYTES (image cache subsystem
    deleted)
  - Drop VAPID_PRIVATE_KEY, VAPID_PUBLIC_KEY, VAPID_CLAIMS_SUB (push
    deleted in phase 8)
  - Drop VOICE_ENABLED, STT_BACKEND, STT_MODEL, TTS_BACKEND (voice
    deleted in phase 8)
  - Drop Config.validate() rules for those keys

Image cache deletion:
  - services/images.py, routes/images.py, models/image_cache.py
  - models/__init__.py: drop ImageCache import
  - app.py: drop images_bp registration
  - alembic/versions/0054_drop_image_cache.py: DROP TABLE image_cache

Frontend client.ts orphan exports stripped:
  - getVoiceStatus, getVoiceList, getVoiceLibrary, installVoice,
    uninstallVoice, transcribeAudio, synthesiseSpeech,
    VoiceStatusResult / VoiceEntry / VoiceLibraryEntry types
  - getJournalConfig, saveJournalConfig, getJournalToday/Day/Days,
    triggerJournalPrep, runJournalCurator, listPendingActions,
    approvePendingAction, rejectPendingAction, listJournalMoments,
    updateJournalMoment, deleteJournalMoment, geocodeAddress
  - JournalConfig / JournalLocation / JournalConversation /
    JournalMessage / JournalDayPayload / JournalMoment /
    CuratorRunResult / PendingCuratorAction types
  - consolidateProfile, clearProfileObservations, listProfileObservations
  - ProfileObservationEntry, learned_summary/observations_* fields on
    UserProfile
  - consolidateTask (cascading update to TaskEditorView)
  - getFableMcpInfo, getNewsItems, GetNewsItemsParams, NewsItem import

TaskEditorView:
  - Drop the auto-summary banner + Re-consolidate button
  - Drop isBodyAutoMaintained gate (editor is always user-controlled now)
  - Drop reconsolidate function + reconsolidating ref

SettingsView:
  - profile ref no longer initialises learned_summary /
    observations_count / observations_updated_at (those fields are
    gone from UserProfile type)

Surviving frontend composables/components flagged for likely future
cleanup but not deleted in this commit (no compile errors, just
unreferenced after Phase 7-8):
  - useAssist, useFloatingAssist, useTagSuggestions, useVad,
    useListenMode, useOnnxPreloader (composables)
  - WorkspaceNoteEditor, WorkspaceTaskPanel, WeatherCard, InlineAssistPanel
    (components)
  - api/client.ts still references /api/notes/assist/* and
    /api/notes/suggest-tags via useAssist + useTagSuggestions — those
    endpoints 404 now but no caller hits them; dead at runtime, harmless.

Compose stack collapses to two services: `app` + `db`. No Ollama, no
voice models, no fable-mcp wheel build. First-boot install reduces to:
  docker compose up -d
  → visit web UI → register → Settings → MCP Access → copy snippet
  → claude mcp add … → done.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/fabledassistant/config.py

@@ -22,30 +22,11 @@ class Config:
         "DATABASE_URL_FILE",
         "postgresql+asyncpg://fabled:fabled@localhost:5432/fabledassistant",
     )
-    OLLAMA_URL: str = os.environ.get("OLLAMA_URL", "http://localhost:11434")
-    OLLAMA_MODEL: str = os.environ.get("OLLAMA_MODEL", "qwen3:latest")
-    # Lightweight model for background tasks (title generation, tag suggestions,
-    # project summaries). Using a separate model keeps the
-    # main model's KV cache intact between user messages, enabling prefix cache hits.
-    OLLAMA_BACKGROUND_MODEL: str = os.environ.get("OLLAMA_BACKGROUND_MODEL", "gemma3:4b")
-    # Ollama keep_alive — how long a model stays resident in VRAM after its last
-    # request. Main model gets a longer window since it's used interactively;
-    # the background model is called sporadically and doesn't need to camp VRAM.
-    # Format matches Ollama's duration strings: "30m", "10m", "1h", "0s", "-1" (forever).
-    OLLAMA_KEEP_ALIVE_MAIN: str = os.environ.get("OLLAMA_KEEP_ALIVE_MAIN", "30m")
-    OLLAMA_KEEP_ALIVE_BACKGROUND: str = os.environ.get("OLLAMA_KEEP_ALIVE_BACKGROUND", "10m")
-    # KV cache context window for generation. Keep this as small as practical —
-    # a larger context forces more KV cache into CPU RAM, drastically slowing prefill.
-    # 16384 covers ~30+ message conversations with our system prompt comfortably.
-    OLLAMA_NUM_CTX: int = int(os.environ.get("OLLAMA_NUM_CTX", "16384"))
     SECRET_KEY: str = _read_secret("SECRET_KEY", "SECRET_KEY_FILE", "dev-secret-change-me")
     SECURE_COOKIES: bool = os.environ.get("SECURE_COOKIES", "").lower() in ("1", "true", "yes")
     LOG_LEVEL: str = os.environ.get("LOG_LEVEL", "INFO")
     LOG_RETENTION_DAYS: int = int(os.environ.get("LOG_RETENTION_DAYS", "90"))
 
-    # Embedding model for semantic note search (served by Ollama)
-    EMBEDDING_MODEL: str = os.environ.get("EMBEDDING_MODEL", "nomic-embed-text")
-
     # SMTP defaults (overridden by DB settings when configured via admin UI)
     SMTP_HOST: str = os.environ.get("SMTP_HOST", "")
     SMTP_PORT: int = int(os.environ.get("SMTP_PORT", "587"))
@@ -64,25 +45,11 @@ class Config:
     OIDC_SCOPES: str = os.environ.get("OIDC_SCOPES", "openid profile email")
     LOCAL_AUTH_ENABLED: bool = os.environ.get("LOCAL_AUTH_ENABLED", "true").lower() not in ("0", "false", "no")
 
-    # SearXNG web search (external instance)
+    # SearXNG web search (external instance). Currently only surfaced via
+    # /api/settings/search for the Integrations tab's status indicator —
+    # the MCP layer doesn't proxy web search (Claude has its own).
     SEARXNG_URL: str = os.environ.get("SEARXNG_URL", "")
 
-    # Image cache — images fetched from the web are stored here and served locally
-    IMAGE_CACHE_DIR: str = os.environ.get("IMAGE_CACHE_DIR", "/data/images")
-    # Maximum size of a single image to cache (default 5 MB)
-    IMAGE_MAX_BYTES: int = int(os.environ.get("IMAGE_MAX_BYTES", str(5 * 1024 * 1024)))
-
-    # VAPID keys for browser push notifications
-    VAPID_PRIVATE_KEY: str = os.environ.get("VAPID_PRIVATE_KEY", "")
-    VAPID_PUBLIC_KEY: str = os.environ.get("VAPID_PUBLIC_KEY", "")
-    VAPID_CLAIMS_SUB: str = os.environ.get("VAPID_CLAIMS_SUB", "mailto:admin@fabledassistant.local")
-
-    # Voice (Speech-to-Speech) feature
-    VOICE_ENABLED: bool = os.environ.get("VOICE_ENABLED", "").lower() in ("1", "true", "yes")
-    STT_BACKEND: str = os.environ.get("STT_BACKEND", "faster-whisper")
-    STT_MODEL: str = os.environ.get("STT_MODEL", "base.en")
-    TTS_BACKEND: str = os.environ.get("TTS_BACKEND", "kokoro")
-
     @classmethod
     def oidc_enabled(cls) -> bool:
         return bool(cls.OIDC_ISSUER and cls.OIDC_CLIENT_ID and cls.OIDC_CLIENT_SECRET)
@@ -95,12 +62,8 @@ class Config:
     def validate(cls) -> None:
         """Validate critical config values at startup. Raises ValueError on misconfiguration."""
         errors: list[str] = []
-        if cls.OLLAMA_NUM_CTX < 512:
-            errors.append(f"OLLAMA_NUM_CTX={cls.OLLAMA_NUM_CTX} is too small (minimum 512)")
         if cls.LOG_RETENTION_DAYS < 1:
             errors.append(f"LOG_RETENTION_DAYS={cls.LOG_RETENTION_DAYS} must be >= 1")
-        if cls.IMAGE_MAX_BYTES < 1024:
-            errors.append(f"IMAGE_MAX_BYTES={cls.IMAGE_MAX_BYTES} must be >= 1024")
         if not (1 <= cls.SMTP_PORT <= 65535):
             errors.append(f"SMTP_PORT={cls.SMTP_PORT} must be between 1 and 65535")
         if cls.oidc_enabled() and not cls.BASE_URL.startswith(("http://", "https://")):
@@ -110,11 +73,5 @@ class Config:
                 "SECRET_KEY is set to the insecure default but SECURE_COOKIES=true indicates "
                 "a production deployment. Set SECRET_KEY or SECRET_KEY_FILE before starting."
             )
-        _valid_stt_models = {"tiny.en", "base.en", "small.en", "medium.en"}
-        if cls.VOICE_ENABLED and cls.STT_MODEL not in _valid_stt_models:
-            errors.append(
-                f"STT_MODEL='{cls.STT_MODEL}' is not supported. "
-                f"Valid values: {', '.join(sorted(_valid_stt_models))}"
-            )
         if errors:
             raise ValueError("Configuration errors:\n" + "\n".join(f"  - {e}" for e in errors))