1. Photo Gallery:
- New `images` JSONB array field in locations schema
- POST /locations/:id/images endpoint to add photos (auth required)
- Gallery with thumbnail strip and image counter on detail page
- Any authenticated user can add photos to any location
- "Add photo" button inline with thumbnails
2. Nearby Locations:
- GET /locations/:id/nearby endpoint with Haversine distance query
- Configurable radius (default 2km, max 10km)
- Returns up to 5 nearby locations sorted by distance
- Horizontal scroll card strip on detail page showing distance
3. Search Suggestions + History:
- GET /locations/suggestions endpoint (prefix matching, fast)
- Search history stored in localStorage (max 8 entries)
- Empty search shows recent history with clock icon
- Selected locations automatically saved to history
- Falls back to full-text search if no prefix matches
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1. Owner tracking (createdBy):
- Add createdBy field to locations schema
- Set createdBy to userId on location creation
- Only owners can edit/delete their own locations
- Seed/unowned locations remain editable by anyone
2. Edit/Delete UI:
- Edit button + full edit form at /locations/:id/edit
- Delete button with confirmation dialog on detail page
- Both only visible to the location owner
- ForbiddenException (403) if non-owner tries to modify
3. Pagination:
- Backend returns paginated results (page, limit, total, totalPages)
- Frontend "Load more" button for infinite scroll
- Category filter reloads from API with server-side filtering
- Default 20 items per page, max 100
Tests updated: 36 tests passing (5 new for ownership + pagination).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Wire mana-llm service into the monitoring stack:
Prometheus (docker/prometheus/prometheus.yml):
- Add mana-llm scrape job (port 3025, 15s interval)
- Include mana-llm in ServiceDown alert expression
Alerts (docker/prometheus/alerts.yml):
- New llm_alerts group with 4 rules:
- LLMServiceDown: mana-llm down > 1 min (critical)
- LLMHighErrorRate: > 10% errors for 5 min (warning)
- OllamaProviderDown: > 50% requests via Google fallback (warning)
- LLMSlowResponses: p95 > 30s for 5 min (warning)
Grafana Dashboard (docker/grafana/dashboards/mana-llm.json):
- 6 stat panels: status, req/min, error rate, fallback rate, latency, tokens/min
- Requests by Provider (stacked area: Ollama vs Google vs OpenRouter)
- Tokens by Type (prompt vs completion)
- Latency Percentiles (p50, p90, p99)
- Latency by Provider comparison
- Requests by Model breakdown
- Errors by Type
- Google Fallback Rate over time (with threshold coloring)
- Provider Distribution pie chart (24h)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The auth service now depends on @manacore/shared-llm but it was missing
from the Docker build context, causing build failures.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Lazy loading for location card images
- Category filter pills now show count (e.g. "Restaurants (3)")
- Better empty states with category-specific messages and emoji
- Share button on detail page (Web Share API with clipboard fallback)
- "On map" and "Directions" (Google Maps) buttons on detail page
- Geolocation "locate me" button on map page with user marker
- Image URL retry button on add form when image fails to load
- i18n keys for all new features (DE/EN)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Migrate matrix-project-doc-bot from raw fetch to @manacore/shared-llm
and remove the unused openai npm package. The bot was already using
mana-llm and mana-stt (not OpenAI directly), but the code still had
raw fetch calls and the openai package installed.
Changes:
- generation.service.ts: raw fetch → llm.chat() via LlmClientService
- app.module.ts: add LlmModule.forRootAsync()
- Remove openai dependency (was unused in code)
- Update CLAUDE.md: document actual AI stack (mana-llm + mana-stt)
- Update TECH_STACK_INDEPENDENCE.md: mark Prio 1-3 as completed
- Prio 1: Picture App → mana-image-gen ✅
- Prio 2: Project Doc Bot → Ollama + mana-stt ✅
- Prio 3: All LLM calls via mana-llm ✅
- Self-hosted percentage: 75% → ~80%
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Rename content collection, pages, and routes from "audits" to
"manascore" for clearer branding. Add comprehensive about page
explaining the 8 scoring categories, weights, and methodology.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add LocalImageGenService that routes to the self-hosted FLUX.2 klein
model on the Mac Mini, eliminating Replicate API dependency for basic
image generation.
Changes:
- LocalImageGenService: wraps mana-image-gen HTTP API (/generate)
with health checking, timeout handling, and GenerationResult compat
- GenerateService: routes to local or Replicate based on model config
(replicateId starting with "local/" → LocalImageGenService)
- Local models always use sync mode (no webhooks needed, ~0.8s)
- Seed: add "FLUX.2 Klein (Lokal)" model with sortOrder -1 (shown first)
- costPerGeneration: 0 (free, runs locally)
- estimatedTimeSeconds: 1
- docker-compose: add IMAGE_GEN_SERVICE_URL env var for picture backend
Replicate remains available for premium models (Seedream, Nano Banana).
Local FLUX.2 klein becomes the default free option.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
All backend containers were sending error reports to glitchtip.mana.how
(via Cloudflare tunnel), creating ~6000 unnecessary external TCP connections
contributing to port exhaustion. Now routes directly to glitchtip:8020
within the Docker network.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Mac Mini had 25k+ TIME_WAIT sockets exhausting the 16k ephemeral port range,
blocking all outgoing TCP connections. Root cause: ~50 health checks at 30s
intervals + n8n automation creating excessive short-lived connections.
- Remove n8n service and volume (no longer needed)
- Increase health check intervals: 30s → 120s (app services), 10s → 30s (infra)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Document complete production readiness checklist, E2E test suites,
and current status for both apps.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace ~70 hardcoded German strings with i18n keys covering calendar
management, view options, reminders, birthdays, and account settings.
All 5 languages (DE/EN/FR/ES/IT).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace ~70 hardcoded German strings with i18n keys in sync and
sharing settings pages. All 5 languages (DE/EN/FR/ES/IT).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds a "Kosten" tab showing all 40+ credit operations across all apps,
grouped by app with category filters (AI, Productivity, Premium) and
color-coded cost badges.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace ~20 hardcoded German toast strings in events, shares, and
external-calendars stores with svelte-i18n keys (5 languages).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Apply shared focusTrap action to 7 app-specific modals for
improved keyboard accessibility and WCAG compliance.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Create reusable Svelte action for focus trapping in modals/dialogs.
Traps Tab/Shift+Tab within element and restores focus on destroy.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Context menus in 12 apps, shared-llm package, onboarding for 16 apps,
todo redesign, QuickInputBar rollout, and dashboard widgets.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add Google Gemini as a fallback provider that activates automatically
when Ollama is overloaded or unavailable, ensuring LLM requests always
succeed even under load.
New provider (src/providers/google.py):
- Full LLMProvider implementation using google-genai SDK
- Chat completions (streaming + non-streaming)
- Vision/multimodal support (base64 images)
- Embeddings via text-embedding-004
- Model mapping: Ollama models → Gemini equivalents
(gemma3:4b → gemini-2.0-flash, llava:7b → gemini-2.0-flash, etc.)
Auto-fallback routing (src/providers/router.py):
- Concurrent request tracking for Ollama (OLLAMA_MAX_CONCURRENT=3)
- When Ollama concurrent > max: route to Google automatically
- When Ollama fails: retry on Google with model mapping
- Health check caching (5s TTL) to avoid hammering Ollama
- Non-Ollama providers (openrouter, groq, together) are never fallback-routed
- Fallback info included in /health endpoint response
New config (src/config.py):
- GOOGLE_API_KEY: enables Google provider
- GOOGLE_DEFAULT_MODEL: default gemini-2.0-flash
- AUTO_FALLBACK_ENABLED: toggle fallback (default: true)
- OLLAMA_MAX_CONCURRENT: concurrent request threshold (default: 3)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add the 3 missing widget components that were registered but not implemented:
- MukkeLibraryWidget: shows library stats (songs, playlists, favorites),
recent songs with artist/duration
- PresiDecksWidget: shows recent presentations with public/private
indicator, descriptions, and update dates
- ContextDocsWidget: shows recent documents grouped by type (text,
context, prompt) with space names, falls back to space list
All widgets follow the established pattern: APP_URLS for links,
WidgetSkeleton/WidgetError for loading/error, auto-retry on failure.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Use shared ContextMenu component on EventCard for quick actions:
edit, duplicate, and delete. Includes i18n for all 5 languages.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add onboarding with feature overview, preference selection, and tips to
Zitare, Mukke, Photos, Planta, SkillTree, and Questions. Insert a new
first "features" info step into all 10 existing onboarding flows so every
app now starts with a core-features overview page.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Use shared ContextMenu component for quick actions: edit, toggle
complete, change priority, move to project, and delete.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Fix todo dev port in APP_URLS (5188, was swapped with inventory 5189)
- Use APP_URLS for all links instead of hardcoded localhost ports
- TasksTodayWidget: add priority dots, subtask progress, label tags,
completed/total counter, clickable task links
- TasksUpcomingWidget: add priority dots, label tags, overdue/today
date highlighting with colored badges, clickable task links
- Extend Task type with labels and subtasks for richer widget display
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Mukke: search songs by title/artist/album via libraryStore
- Matrix: search rooms/contacts, select navigates to chat
- ManaDeck: search decks by title/description
- Planta: search plants by name/species via plantsApi
- Photos: search albums and tags
- Presi: search presentation decks
All with locale-aware syntax highlighting.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Changes in the expanded inline task editor are now saved automatically
with a 500ms debounce. Removes the need to manually click "Speichern".
The delete button remains for intentional destructive actions.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Remove spiral binding holes (ring elements) from notepad container
- Remove vertical red margin line from left side
- Add X close button to onboarding tip so users can dismiss it
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add `locale` prop to all 6 apps using QuickInputBar
(todo, contacts, zitare, citycorners, questions, calendar)
- Enable `deferSearch` on apps with create flow
(contacts, zitare, questions) to match todo behavior
- Pass locale through Calendar's UnifiedBar wrapper
- Questions: default to 'en' locale (English-first app)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Extract applyTaskFilters as pure utility function for testability
- Add 21 tests for task filtering (priority, project, labels, search, combined)
- Add 13 tests for viewStore filter methods (setters, clearFilters, reset)
- All 87 tests passing
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
1. Extract hardcoded German highlight patterns into locale-specific sets
(de, en, fr, it, es). InputBar accepts `locale` or custom
`highlightPatterns` prop, defaulting to German for backward compat.
2. Add visual success feedback after creating: input bar flashes green
with a checkmark icon for 1.2s, confirming the action was successful.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Rewrite TaskListSkeleton to match notepad design (spiral holes, red margin line, cream background)
- Rewrite TaskItemSkeleton as flat rows with border-bottom instead of card style
- Rewrite KanbanColumnSkeleton with glassmorphism and pill-shaped task cards
- Update KanbanBoardSkeleton with matching border-radius and dark mode support
- Group navigation pills (Liste/Kanban/Tags) into PillTabGroup for clear UX distinction from toggle pills (Filter)
- Add Phosphor icon support to PillTabGroup component
- Fix %sveltekit.head% appearing as literal text in production by removing duplicate placeholder from HTML comment
- Disable PWA service worker in dev mode
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
