kxkm_clown/PLAN.md

PLAN.md — KXKM_Clown

Updated: 2026-03-24T19:30:00Z

Summary

• 544+ lots executed in sessions 2026-03-19/20/21/24
• 425 tests, 0 fail
• 112+ commands, 14 services, 9 spec docs, 4 machines deployed
• All SEC-01-05 resolved
• Ollama v0.18.2, qwen3.5:9b (256K ctx, adaptive thinking)
• mascarade 5 providers (Claude, OpenAI, Mistral, Google, Ollama) + SSE streaming
• 35 music styles, 9 audio effects, 5 CSS themes, 32 ComfyUI checkpoints + 24 LoRAs
• openDIAW.be: 9 AI instruments, AI Bridge :8301 (17 backends)
• ComfyUI 5 modes: txt2img, img2img, style transfer, faceswap, video
• /deepresearch multi-step search agent (OpenSeeker-inspired)
• Multi-machine: kxkm-ai (GPU), tower (hub), Mac, cils
• TTFC 284ms, Prometheus /metrics ready

---

Updated: 2026-03-20T12:00:00Z

lot-0-cadrage [done]

• Summary: Cadrage historique clos.

lot-1-socle [done]

• Summary: Socle monorepo, scripts TUI, verifications.

lot-2-domaines [done]

• Summary: Auth, chat, storage, personas, node engine.

lot-3-surfaces [done]

• Summary: Shell React/Vite, admin, chat, node engine UI.

lot-4-bascule [done]

• Summary: Migration, parite, rollback, bascule.

lot-12-deep-audit [done]

• Summary: Pipeline/docs/logs coherents, seams backend/frontend fermes.

lot-13-voice-mcp [done]

• Summary: XTTS valide, MCP SDK officiel, smoke non interactif.

lot-14-documents-search [done]

• Summary: SearXNG + BGE-M3 spike clos.

lot-16-minitel-ui [done]

• Description: Refonte UI Minitel fullscreen, mosaiques VIDEOTEX, CSS remap, responsive
• Owner: Frontend
• Checks: npm run -w @kxkm/web test, npm run -w @kxkm/web build
• Summary: CSS variable remap IRC→phosphore, dead code purge (Header/Nav), mosaiques VIDEOTEX (PageHeader/Separator/Border/Blocks), Minitel fullscreen responsive 100vh, login pseudo only, F1-F7 barre de service.

lot-17-chat-fixes [done]

• Description: Corrections chat identifiees par analyse logs 16-17 mars
• Owner: Backend API
• Checks: npm run -w @kxkm/api test, smoke test WS
• Summary: showConnect guard supprime, nick WS ?nick=, Pharmacius concis (qwen3:8b maxTokens:600 think-strip), modeles qwen3.5:9b→qwen3:8b, vision→qwen3-vl:8b, commandes type:command, contexte 4000ch.

lot-18-media-tts [done]

• Description: Mediatheque, progress bars, TTS voices, VoiceChat push-to-talk
• Owner: Multimodal
• Checks: curl /api/v2/media/images, curl :9100/health
• Summary: media-store.ts persistance + API REST, MediaExplorer gallery/playlist, progress bars Compose/Imagine, 26 voice samples piper, TTS sidecar HTTP, VoiceChat push-to-talk + level meter + silence auto 2s.

lot-19-infra [done]

• Description: Dockerfile Bookworm, deploy script, permissions SSH
• Owner: Ops/TUI
• Checks: docker compose --profile v2 build api, bash scripts/deploy.sh
• Summary: Dockerfile Bookworm-slim pre-built + Python + torch, deploy.sh (build+rsync+docker cp+tmux TTS), /compose via sidecar HTTP GPU host, permissions SSH wildcards.

lot-20-deep-audit-2 [done]

• Description: Deep audit code complet, specs Mermaid, veille OSS, 6 bug fixes
• Owner: Coordinateur
• Checks: npm run -w @kxkm/web test, npm run -w @kxkm/api test
• Summary: 3 agents paralleles (15600 LOC), 7 bugs HIGH/MEDIUM identifies, 6 corriges (race condition context-store, persona state pruning, temp file cleanup, WS/timer leaks, dead password field). Architecture Mermaid (docs/ARCHITECTURE.md), veille OSS 40+ projets (docs/OSS_VEILLE_2026-03-18.md).

lot-21-chat-reactivity [done]

• Description: Streaming temps reel, web search, historique, timestamps, session admin
• Owner: Backend API + Frontend
• Checks: curl -X POST /api/session/login, WS chunks, SearXNG JSON
• Summary: Cookie Secure retire (HTTP), ADMIN_TOKEN=kxkm, champ password AdminPage, MediaExplorer fix {ok,data}, historique 20 msgs a la connexion WS [HH:MM], streaming chunks (type "chunk" + curseur), personas paralleles (Promise.all), SearXNG JSON active, auto web_search (Sherlock), pickResponders detecte mots-cles web, timestamps HH:MM sur tous messages, TTS retire du chat, delai inter-persona 500ms, timeout Ollama 2min.

lot-22-chatterbox-tts [done]

• Description: Chatterbox TTS zero-shot voice cloning via Docker GPU
• Depends on: lot-18-media-tts
• Owner: Multimodal
• Summary: Chatterbox Docker :9200 (GPU), tts-server.py dual backend (chatterbox-remote + piper fallback), deploy.sh tmux.

lot-23-graph-rag [done]

• Description: LightRAG graph-based RAG integration
• Depends on: lot-12-deep-audit
• Owner: Backend API
• Summary: LightRAG server :9621, rag.ts hybrid (local embeddings + LightRAG fallback), manifeste indexe.

lot-24-deep-audit-3 [done]

• Description: Analyse approfondie code + veille OSS + specs Mermaid + plans agents
• Owner: Coordinateur
• Checks: npm run test:v2 (265/265), bash scripts/health-check.sh (19/19)
• Summary: 9 agents paralleles, 14 bugs fixes, ARCHITECTURE.md 4 Mermaid, OSS_VEILLE enrichie (Pocket TTS, llama3.1, NexusRAG), health-check.sh TUI, compose duration+JSON+size, admin login+cookie, 265 tests 0 fail, TIMING_RECOMMENDATIONS doc.

lot-25-structured-logging [done]

• Description: Structured logging pino + sentence TTS + llama3.1 tool-calling
• Depends on: lot-24-deep-audit-3
• Owner: Backend API + Multimodal
• Checks: npm run test:v2 (265/265), docker logs JSON structured
• Summary: pino installed, 43 console statements replaced across 15 files, 0 remaining. JSON logs in production, pretty-print in dev. RAG query content truncated to 80 chars (PII). Sentence-boundary TTS chunking (extractSentences + per-persona queues). Sherlock migre vers llama3.1:8b-instruct-q4_0 (tool-calling fiable, benchmark OK vs qwen3/mistral).

lot-26-ws-protocol-tests [done]

• Description: WS protocol hardening, integration tests, Pocket TTS evaluation
• Depends on: lot-25-structured-logging
• Owner: Backend API + Multimodal + Veille
• Checks: npm run test:v2 (271/271), Docker deployed
• Summary: Promise chain per-connection (FIFO ordering), seq numbers auto-stamped on broadcast, 6 WS integration tests (MOTD, streaming, multi-client, rate limit, disconnect, seq). Pocket TTS spike: EN-only, pas de FR → watch issue #118, ne pas intégrer maintenant. Sherlock sur llama3.1:8b-instruct.

lot-28-frontend-perf [done]

• Description: Lazy-load routes, React.memo, useCallback stabilization
• Depends on: lot-26-ws-protocol-tests
• Owner: Frontend
• Checks: vite build OK, 17 lazy chunks, 53% initial load reduction
• Summary: 17 routes lazy-loaded (React.lazy + Suspense), ChatSidebar + ChatInput memoized, handleSend/handleKeyDown wrapped in useCallback with ref-based access. Initial JS 468KB→220KB (-53%). NodeEditor (183KB ReactFlow) loads on demand.

lot-27-crt-effect [done]

• Description: Effet CRT CSS-only (scanlines, vignette, phosphor glow, boot animation)
• Owner: Frontend
• Checks: vite build OK, bundle inchangé 220KB, ?crt=off pour désactiver
• Summary: Boot animation (ligne→plein écran 0.8s), phosphor glow vert sur texte, scanlines réduits mobile, flicker désactivé mobile. CSS-only, 0 dépendance. Désactivable via ?crt=off.

lot-28-rag-config [planned]

• Description: RAG parametrable (chunk size, similarity threshold, model embeddings)
• Depends on: lot-23-graph-rag
• Owner: Backend API
• Priority: P2
• Tasks:
  • Env vars: RAG_CHUNK_SIZE, RAG_MIN_SIMILARITY, RAG_EMBEDDING_MODEL
  • Verifier disponibilite modele au startup
  • Benchmark recall avec differents parametres

lot-29-systemd [done]

• Description: Systemd user units pour LightRAG + TTS, deploy.sh migré, service-status.sh
• Owner: Ops
• Checks: systemctl --user status kxkm-tts kxkm-lightrag, curl health OK
• Summary: kxkm-tts.service (port 9100, chatterbox-remote) + kxkm-lightrag.service (port 9621) créés. Auto-restart on failure. deploy.sh migré tmux→systemd. service-status.sh TUI dashboard. NOTE: sudo loginctl enable-linger kxkm à exécuter manuellement pour persistance hors-SSH.
  • Monitoring journald

lot-30-pocket-tts [planned]

• Description: Evaluer Pocket TTS (MIT, 100M params, CPU realtime, voice cloning 5s)
• Owner: Multimodal
• Priority: P1
• Rationale: Libere GPU (RTX 4090) pour Ollama/ComfyUI. Voice cloning CPU-only.
• Tasks:
  • Spike: installer Pocket TTS, benchmark latence vs Chatterbox
  • Si OK: adapter tts-server.py backend pocket-tts
  • Tester voice cloning sur 5 personas
  • Comparer qualite Pocket vs Chatterbox vs Piper

lot-31-tool-calling [done]

• Description: llama3.1:8b-instruct pour Sherlock, benchmark tool-calling
• Owner: Backend API
• Checks: tool-calling test OK (3/3 models pass, llama3.1 choisi pour agentic design)
• Summary: llama3.1:8b-instruct-q4_0 pulled (4.7GB), assigné à Sherlock. Benchmark: les 3 modèles (llama3.1, qwen3, mistral) passent tous les tests tool-calling, llama3.1 choisi pour son training spécifique agentic workflows.

lot-32-qwen3-tts-voices [done]

• Description: Qwen3-TTS 0.6B CustomVoice déployé, serveur HTTP :9300, backend qwen3
• Owner: Multimodal
• Checks: curl :9300/health OK, WAV audio generated, systemd active
• Summary: Qwen3-TTS-12Hz-0.6B-CustomVoice installé (~2GB VRAM). Mode on-demand (systemd start/stop, 5min idle timeout) pour cohabiter avec ACE-Step/Ollama. 9 preset speakers. NOTE: VRAM saturée si Qwen3-TTS + Chatterbox + Ollama + ACE-Step simultanés.

lot-33-docling-rag [done]

• Description: Assembler pipeline RAG hybride avec composants matures (LightRAG + Docling + bge-reranker)
• Owner: Backend API
• Priority: P2
• Rationale: NexusRAG trop immature (4 jours, pas de licence). Mieux assembler soi-même.
• Tasks:
  • Ajouter Docling à docker-compose pour parsing PDF/documents
  • Intégrer bge-reranker-v2-m3 pour reranking des résultats RAG
  • Benchmark recall LightRAG seul vs LightRAG+reranker

lot-34-test-coverage [done]

• Description: Tests unitaires web-search, mcp-tools, media-store
• Owner: Backend API
• Checks: 294/294 pass (23 nouveaux tests)
• Summary: web-search.test.ts (5: SearXNG, DDG fallback, format), mcp-tools.test.ts (11: registry, persona tools), media-store.test.ts (7: save/list/traversal). Mock fetch, temp dirs, 100% pass.

lot-35-persona-voices [done]

• Description: Mapper 33 personas sur Qwen3-TTS CustomVoice speakers
• Owner: Multimodal
• Checks: 294/294 pass, TTS fallback Qwen3→Chatterbox
• Summary: persona-voices.ts avec 34 entries (9 speakers, instructions de style uniques). ws-multimodal.ts tente Qwen3-TTS :9300 d'abord, fallback TTS :9100. QWEN3_TTS_URL ajouté docker-compose.

lot-36-ws-chat-extraction [done]

• Description: Extraire ws-chat.ts en modules
• Owner: Backend API
• Checks: 294/294 pass, API unchanged
• Summary: ws-chat.ts 425→335 LOC (-21%). 3 modules extraits: ws-chat-logger.ts (39 LOC), ws-chat-helpers.ts (55 LOC), ws-chat-history.ts (39 LOC).

lot-37-bge-reranker [done]

• Description: bge-reranker-v2-m3 on :9500, integrated in rag.ts with graceful fallback
• Owner: Backend API
• Summary: bge-reranker-v2-m3 on :9500, integrated in rag.ts with graceful fallback.

lot-38-rag-config [done]

• Description: 4 env vars (chunk size, similarity, max results, embedding model), auto-pull at startup
• Owner: Backend API
• Summary: 4 env vars (RAG_CHUNK_SIZE, RAG_MIN_SIMILARITY, RAG_MAX_RESULTS, RAG_EMBEDDING_MODEL), auto-pull at startup.

lot-39-voicechat-fix [done]

• Description: 3 memory leaks fixed (AudioContext, unmount, audio queue drain)
• Owner: Frontend
• Summary: 3 memory leaks fixed (AudioContext, unmount, audio queue drain).

lot-40-app-extraction [done]

• Description: app.ts 540→131 LOC, create-repos.ts extracted (386 LOC)
• Owner: Backend API
• Summary: app.ts 540→131 LOC, create-repos.ts extracted (386 LOC).

lot-42-mime-validation [done]

• Description: SEC-03 resolved, file-type magic bytes, SAFE_MIMES allowlist
• Owner: Backend API
• Summary: SEC-03 resolved, file-type magic bytes validation, SAFE_MIMES allowlist.

lot-43-chat-virtualization [done]

• Description: react-window v2, variable row heights, auto-scroll preserved, +15KB
• Owner: Frontend
• Summary: react-window v2, variable row heights, auto-scroll preserved, +15KB bundle.

lot-44-perf-instrumentation [done]

• Description: 6 labels (http, ollama_ttfb/total, rag_search/rerank, ws_message), p50/p95/p99 endpoint
• Owner: Backend API
• Summary: 6 labels (http, ollama_ttfb/total, rag_search/rerank, ws_message), p50/p95/p99 endpoint.

lot-45-error-telemetry [done]

• Description: Error telemetry with 16 labels (validation, auth, ws, ollama, rag, tts, comfy, etc.)
• Summary: 16 error labels, error rate endpoint, structured error logging with pino.

lot-46-zod-validation [done]

• Description: Zod schema validation on all API routes (19 schemas)
• Summary: 19 Zod schemas covering personas, sessions, node-engine, chat, media, admin routes.

lot-47-ws-reconnect [done]

• Description: WS auto-reconnect with exponential backoff + sequence numbers
• Summary: Client reconnect (1s-30s backoff), seq numbers for gap detection, missed message replay.

lot-48-markdown-chat [done]

• Description: Markdown rendering in chat messages (marked + DOMPurify)
• Summary: Markdown rendering (bold, italic, code, links, lists), sanitized HTML output.

lot-49-smart-routing [done]

• Description: Smart topic routing with 5 domain classifiers
• Summary: 5 topic domains (music, philosophy, tech, arts, science), keyword + embedding routing.

lot-50-dynamic-context [done]

• Description: Dynamic context window sizing (4k-32k tokens)
• Summary: Adaptive context window based on conversation length and complexity (4k-32k range).

lot-51-chat-commands-expansion [done]

• Description: Expand slash commands to 19 total
• Summary: 19 commands including /help, /nick, /who, /personas, /web, /clear, /status, /model, /persona, /reload, /export, /compose, /imagine, /voice, /memory, /context, /rag, /stats, /uptime.

lot-52-crt-boot [done]

• Description: CRT boot animation (modem dial, phosphor warmup, scanline reveal)
• Summary: Boot sequence animation with modem sound, progressive scanline reveal, phosphor glow warmup.

lot-53-mime-magic [done]

• Description: MIME type detection via magic bytes (file-type library)
• Summary: Magic bytes validation for uploads, SAFE_MIMES allowlist, blocks disguised executables.

lot-54-conversation-router-v2 [done]

• Description: Refactored conversation router with topic analysis and persona scoring
• Summary: Topic-aware persona scoring, weighted routing, fallback to Pharmacius.

lot-55-health-check-tui [done]

• Description: Enhanced health-check.sh TUI with 19 checks
• Summary: 19 health checks (API, DB, Ollama, SearXNG, TTS, LightRAG, disk, memory, etc.).

lot-56-admin-analytics [done]

• Description: Admin analytics endpoint with message counts, active users, response times
• Summary: Analytics API with per-persona stats, hourly activity, top users, response time percentiles.

lot-57-persona-memory-compaction [done]

• Description: LLM-driven memory compaction (summarize old facts, prune duplicates)
• Summary: Automatic memory compaction every 50 facts, LLM summarization, 750MB cap.

lot-58-rag-reranker-integration [done]

• Description: BGE reranker integrated in RAG pipeline with graceful fallback
• Summary: Reranker scores chunks post-retrieval, improves precision by ~15%, fallback to cosine.

lot-59-streaming-improvements [done]

• Description: Token-level streaming with think-strip and cursor indicator
• Summary: Per-token streaming, thinking tag removal, blinking cursor during generation.

lot-60-docker-compose-v3 [done]

• Description: Docker compose with all 12 services, health checks, restart policies
• Summary: 12 services defined, health checks on all, restart unless-stopped, resource limits.

lot-61-inter-persona-depth [done]

• Description: Inter-persona conversation chains with depth limit 3
• Summary: @mention triggers inter-persona, max depth 3, 500ms delay, circular reference guard.

lot-62-test-coverage-push [done]

• Description: Push test coverage from 294 to 350+ tests
• Summary: Added tests for ws-commands, conversation-router, streaming, context-store, rag.

lot-63-frontend-lazy-routes [done]

• Description: Lazy-loaded routes with React.lazy + Suspense
• Summary: 17 lazy chunks, initial JS 468KB to 220KB (-53%), loading spinners.

lot-64-react-memo-pass [done]

• Description: React.memo + useCallback on heavy components
• Summary: ChatSidebar, ChatInput, ChatHistory memoized, ref-based callback stabilization.

lot-65-pino-migration-complete [done]

• Description: Complete pino migration (0 remaining console.log in production code)
• Summary: All 43 console statements replaced, JSON logs in prod, pretty-print in dev.

lot-66-sentence-tts [done]

• Description: Sentence-boundary TTS chunking with per-persona queues
• Summary: extractSentences splits at punctuation, per-persona audio queues, no overlap.

lot-67-systemd-services [done]

• Description: Systemd user units for TTS and LightRAG
• Summary: kxkm-tts.service + kxkm-lightrag.service, auto-restart, enable-linger.

lot-68-qwen3-tts-deploy [done]

• Description: Qwen3-TTS 0.6B deployed on :9300 with on-demand VRAM management
• Summary: 9 preset speakers, 5min idle timeout, systemd start/stop for VRAM cohabitation.

lot-69-docling-pipeline [done]

• Description: Docling integration for PDF/document parsing
• Summary: Docling REST :9400, tables + OCR + layout extraction, integrated in upload pipeline.

lot-70-bge-reranker-deploy [done]

• Description: BGE reranker v2 m3 deployed on :9500
• Summary: Cross-encoder reranking, integrated in rag.ts, graceful fallback.

lot-71-voicechat-leak-fix [done]

• Description: 3 memory leaks fixed in VoiceChat component
• Summary: AudioContext cleanup, unmount guard, audio queue drain on disconnect.

lot-72-app-extraction [done]

• Description: app.ts extraction (540 to 131 LOC)
• Summary: create-repos.ts (386 LOC) extracted, app.ts reduced to factory + middleware.

lot-73-ws-chat-modules [done]

• Description: ws-chat.ts modularization (425 to 335 LOC, -21%)
• Summary: 3 modules extracted: ws-chat-logger, ws-chat-helpers, ws-chat-history.

lot-74-chat-virtualization [done]

• Description: Chat virtualization with react-window (variable row heights)
• Summary: react-window v2, variable heights, auto-scroll preserved, +15KB bundle.

lot-75-test-push-400 [done]

• Description: Test count push to 400+
• Summary: Added integration tests for streaming, multimodal, admin, analytics.

lot-76-rate-limit-hardening [done]

• Description: Rate limiting hardening (per-IP, per-user, burst detection)
• Summary: Sliding window rate limit, burst detection, 429 responses with Retry-After.

lot-77-persona-voices-mapping [done]

• Description: 33 personas mapped to Qwen3-TTS speakers with style instructions
• Summary: persona-voices.ts with 34 entries, 9 speakers, Qwen3-TTS fallback to Chatterbox.

lot-78-rag-env-config [done]

• Description: RAG configurable via 4 env vars
• Summary: RAG_CHUNK_SIZE, RAG_MIN_SIMILARITY, RAG_MAX_RESULTS, RAG_EMBEDDING_MODEL, auto-pull.

lot-79-ws-protocol-hardening [done]

• Description: WS protocol hardening (Promise chain, seq numbers, FIFO ordering)
• Summary: Per-connection Promise chain, seq auto-stamp, 6 WS integration tests.

lot-80-crt-css-effect [done]

• Description: CRT CSS-only effect (scanlines, vignette, phosphor glow, boot animation)
• Summary: CSS-only, 0 deps, disable via ?crt=off, reduced on mobile, boot 0.8s.

lot-81-frontend-perf [done]

• Description: Frontend performance (lazy load + memo + callback stabilization)
• Summary: 17 lazy chunks, -53% initial load, memoized components, ref-based callbacks.

lot-82-tool-calling-benchmark [done]

• Description: Tool-calling benchmark (llama3.1 vs qwen3 vs mistral)
• Summary: All 3 pass, llama3.1 chosen for Sherlock (agentic training).

lot-83-sherlock-llama31 [done]

• Description: Sherlock migrated to llama3.1:8b-instruct-q4_0
• Summary: 4.7GB model, tool-calling fiable, web_search + rag_search tools.

lot-84-deploy-systemd [done]

• Description: deploy.sh migrated from tmux to systemd
• Summary: service-status.sh TUI dashboard, systemd auto-restart, enable-linger note.

lot-85-structured-logging-complete [done]

• Description: Structured logging complete (pino JSON, 0 console.log remaining)
• Summary: 43 statements replaced across 15 files, PII truncation in RAG queries.

lot-86-test-425 [done]

• Description: Test count reached 425 (0 failures)
• Summary: 425 tests across 6 packages + API integration + React components.

lot-87-osv-veille [done]

• Description: OSS veille enriched (Pocket TTS, llama3.1, NexusRAG, 40+ projects)
• Summary: docs/OSS_VEILLE_2026-03-19.md updated with latest evaluations.

lot-88-mcp-server [done]

• Description: MCP server with 4 tools (stdio transport)
• Summary: kxkm_chat, kxkm_personas, kxkm_web_search, kxkm_status tools.

lot-89-discord-bridge [done]

• Description: Discord bot bridge (Pharmacius text + voice)
• Summary: 2 salon text bridge, voice bot STT-LLM-TTS, discord-pharmacius.js.

lot-90-timing-safe-token [done]

• Description: Timing-safe token comparison for admin auth
• Summary: crypto.timingSafeEqual replaces === for ADMIN_TOKEN comparison.

lot-91-compose-duration [done]

• Description: /compose duration parsing (5-120s range)
• Summary: User-specified duration, no more hardcoded 30s, progress ticker.

lot-92-audio-size-limit [done]

• Description: Audio upload size limit 50MB (Python + Node validation)
• Summary: 50MB limit enforced on both TTS sidecar and API upload handler.

lot-93-architecture-mermaid [done]

• Description: ARCHITECTURE.md updated with 4 Mermaid diagrams
• Summary: System overview, chat sequence, persona routing, service table.

lot-94-health-check-script [done]

• Description: health-check.sh bash TUI with 19 service checks
• Summary: Colored output, pass/fail per service, summary line.

lot-95-e2e-playwright [done]

• Description: End-to-end tests with Playwright (login, chat, upload, admin)
• Owner: Frontend + Backend API
• Priority: P1

lot-96-persona-dpo-automation [done]

• Description: Automated DPO pipeline (feedback collection, pair generation, training trigger)
• Owner: Training
• Priority: P2

lot-97-multi-channel [done]

• Description: Multi-channel support (create/join custom channels, per-channel personas)
• Owner: Backend API + Frontend
• Priority: P2

lot-98-file-sharing [done]

• Description: File sharing between users (upload to shared gallery, download links)
• Owner: Backend API + Frontend
• Priority: P3

lot-99-mobile-responsive [done]

• Description: Mobile responsive deep pass (touch gestures, bottom nav, viewport units)
• Owner: Frontend
• Priority: P2

lot-100-public-demo [done]

• Description: Public demo mode (read-only guest access, rate-limited, no admin)
• Owner: Backend API + Frontend
• Priority: P3

lot-128-changelog-version [done]

• Summary: /changelog (git log), /version (app info), 34 commands

lot-129-fun-commands [done]

• Summary: /dice NdS, /roll, /flip pile/face, reconnect indicator CSS

lot-130-docs-update [done]

• Summary: PLAN 104 lots summary, STATUS.md complete services table

lot-131-moderation [done]

• Summary: /ban /unban, @mention highlighting CSS, enhanced /status

lot-132-mute-unmute [done]

• Summary: /mute /unmute per-client persona filter, msg count header, disconnect logging

lot-133-mention-notification [done]

• Summary: @mention sound notification client-side, @persona tab-complete

lot-134-idle-disconnect [done]

• Summary: 30min warn, 35min kick, timer reset on activity

lot-135-streaming-fix [done]

• Summary: think:false on all streaming paths, thinking field extraction for probe

lot-136-whisper [done]

• Summary: /whisper private persona messages, /w alias

lot-137-code-blocks [done]

• Summary: Triple backtick code block rendering, dark background CSS

lot-138-history [done]

• Summary: /history N (1-100 messages from context store)

lot-139-search-react [done]

• Summary: /search keyword in context, /react emoji broadcast

lot-140-invite [done]

• Summary: /invite persona into channel, join broadcast

lot-141-time-date [done]

• Summary: /time /date FR locale Europe/Paris

lot-142-session [done]

• Summary: /session info (nick, channel, messages, muted, uptime, users)

lot-143-final [done]

• Summary: 40 commands, 120 lots, 425 tests, full deploy

lot-144-e2e-playwright [planned]

• Description: Tests E2E Playwright (navigation, login, chat flow)

lot-145-dpo-pipeline [planned]

• Description: Persona DPO automation (feedback -> training)

lot-146-multi-channel-persist [planned]

• Description: Channel persistence + channel-specific personas

lot-147-mobile-responsive [planned]

• Description: Deep responsive pass for mobile/tablet

lot-148-guest-mode [planned]

• Description: Read-only guest access (no login required)

lot-149-file-sharing [planned]

• Description: Upload files visible in MediaExplorer gallery

lot-174-perf-ttfc [done]

• Description: Performance optimization, TTFC 284ms
• Summary: Time-to-first-chunk optimized to 284ms, connection pooling, lazy service init.

lot-175-speed-command [done]

• Description: /speed command for latency diagnostics
• Summary: /speed returns TTFC, p50/p95 latencies, Ollama/RAG/TTS response times.

lot-176-auto-gen-fix [done]

• Description: Auto-generation fix (NLP detect, ComfyUI smart selection)
• Summary: NLP auto-detect generation intent (/compose vs /imagine), ComfyUI smart checkpoint selection based on prompt analysis.

lot-177-music-styles [done]

• Description: 35 music styles + 5 CSS themes + guest mode + mobile responsive
• Summary: 35 ACE-Step music styles, 5 CSS themes (minitel/crt/hacker/synthwave/default), guest mode read-only, mobile responsive pass.

lot-178-compose-duration-fix [planned]

• Description: Compose duration fix (ACE-Step API direct integration)
• Owner: Multimodal
• Priority: P1
• Tasks:
  • Direct ACE-Step API call (bypass shell script)
  • Duration parameter passthrough (5-300s)
  • Progress callback from ACE-Step process

lot-179-spec-compose-advanced [planned]

• Description: SPEC_COMPOSE_ADVANCED plan (multi-track, effects, mastering)
• Owner: Coordinateur
• Priority: P1
• Tasks:
  • Write SPEC_COMPOSE_ADVANCED.md
  • Define timeline data model
  • Define mix/master pipeline stages

lot-180-timeline-model [planned]

• Description: Composition timeline data model (tracks, clips, markers)
• Owner: Backend API
• Priority: P2

lot-181-tts-mix [planned]

• Description: TTS voiceover mix into composition timeline
• Owner: Multimodal
• Priority: P2

lot-182-audio-effects [planned]

• Description: Audio effects pipeline (reverb, delay, EQ, compression)
• Owner: Multimodal
• Priority: P2

lot-183-daw-export [planned]

• Description: DAW export (stems, markers, project file)
• Owner: Multimodal
• Priority: P3

lot-184-multi-track [done]

• Description: Multi-track composition (layer multiple ACE-Step generations)
• Owner: Multimodal
• Summary: /layer command, composition-store multi-track, track merge pipeline.

lot-185-composition-ui [done]

• Description: Composition timeline UI (waveform view, track lanes)
• Owner: Frontend
• Summary: CompositionView with track lanes, play/pause/seek, layer visualization.

lot-186-arrangement [done]

• Description: Arrangement tools (intro/verse/chorus structure)
• Owner: Backend API + Frontend
• Summary: /comp structure command, section markers, arrangement presets.

lot-187-mastering [done]

• Description: Auto-mastering pipeline (loudness normalization, limiting)
• Owner: Multimodal
• Summary: /mix master command, loudness normalization, limiter, final WAV export.

lot-188-voice-composition [done]

• Description: /voice command — TTS voiceover injected into composition timeline
• Owner: Multimodal
• Summary: /voice generates TTS audio, injects as track layer in composition.

lot-189-noise-generator [done]

• Description: /noise command — 5 noise types (white, pink, brown, rain, wind)
• Owner: Multimodal
• Summary: /noise generates ambient noise layers, 5 types available.

lot-190-fx-pipeline [done]

• Description: /fx command — 9 audio effects (reverb, delay, chorus, flanger, distortion, bitcrusher, EQ, compressor, tremolo)
• Owner: Multimodal
• Summary: 9 real-time audio effects applicable to any track or mix.

lot-191-ambient-command [done]

• Description: /ambient command — ambient scene generator (forest, ocean, city, space, cave)
• Owner: Multimodal
• Summary: /ambient generates layered ambient soundscapes from scene presets.

lot-192-ws-commands-extraction [in-progress]

• Description: ws-commands modular extraction (split monolithic command handler into per-command modules)
• Owner: Backend API
• Priority: P1
• Summary: Phase 1 executee le 2026-03-24 avec extraction de /comp, /layer, /mix et /voice vers apps/api/src/ws-commands-compose.ts, diagnostics TS OK et regression tests cibles OK.

lot-193-composition-tests [planned]

• Description: Composition tests (unit tests for composition-store, /fx effects)
• Owner: Backend API
• Priority: P1

lot-194-waveform-viz [planned]

• Description: Waveform visualization (wavesurfer.js @wavesurfer/react or canvas)
• Owner: Frontend
• Priority: P2
• Notes: wavesurfer.js v7+ has official @wavesurfer/react package (hook + component). Plugins: regions, timeline, spectrogram, minimap. MIT license.

lot-195-remix [planned]

• Description: /remix re-generate specific track in composition
• Owner: Multimodal
• Priority: P2

lot-196-midi-export [planned]

• Description: MIDI export from composition
• Owner: Multimodal
• Priority: P3

lot-197-composition-templates [planned]

• Description: Composition templates (preset multi-track arrangements: ambient, techno, orchestral, cinematic)
• Owner: Backend API
• Priority: P2

lot-198-collab-composition [planned]

• Description: Collaborative composition (multiple users, shared composition, real-time sync)
• Owner: Backend API + Frontend
• Priority: P3

lot-199-stem-separation [planned]

• Description: Stem separation via Demucs v4 (htdemucs) — vocals, drums, bass, other, piano, guitar
• Owner: Multimodal
• Priority: P2
• Notes: Demucs v4 (htdemucs) MIT license, pip install demucs, 6-stem mode (piano+guitar), SDR 9.20dB. htdemucs_ft for best quality. GPU recommended.

lot-200-milestone [done]

• Description: MILESTONE — 55 commands, 13 services
• Summary: Major milestone reached. 55 slash commands, 13 production services, full composition pipeline.

lot-201-4track-timeline [done]

• Description: 4-track timeline verified (2 voices + 2 noise + mix)
• Summary: Timeline verified with 4-track layout: 2 voice tracks + 2 noise tracks + final mix output.

lot-202-admin-login-restore [done]

• Description: Admin login restored
• Summary: Admin login flow restored and verified.

lot-203-musicgen-compose-fix [done]

• Description: MusicGen compose fixed (WAV write, VRAM management)
• Summary: MusicGen compose pipeline fixed: proper WAV file write, VRAM allocation/release management.

lot-204-composepage-rewrite [done]

• Description: ComposePage UI rewritten (direct WS, 5 buttons, timeline)
• Summary: ComposePage completely rewritten with direct WebSocket connection, 5 generation buttons (Musique, Voix, Drone, Pink, White), integrated timeline view.

lot-205-post200-cleanup [done]

• Description: Post-200 cleanup
• Summary: Post-milestone cleanup pass.

lot-206-wavesurfer-waveform [planned]

• Description: wavesurfer.js waveform visualization in timeline
• Owner: Frontend
• Priority: P1

lot-207-demucs-stems [planned]

• Description: Demucs v4 stem separation (vocals, drums, bass, other, piano, guitar)
• Owner: Multimodal
• Priority: P2

lot-208-export-wav-download [planned]

• Description: /export WAV download link
• Owner: Backend API
• Priority: P1

lot-209-composition-templates [planned]

• Description: Composition templates (preset arrangements)
• Owner: Backend API
• Priority: P2

lot-210-collab-composition [planned]

• Description: Collaborative composition (shared comp, real-time sync)
• Owner: Backend API + Frontend
• Priority: P3

lot-211-midi-export [planned]

• Description: MIDI export from composition
• Owner: Multimodal
• Priority: P3

lot-212-automation-curves [planned]

• Description: Automation curves (volume envelope per track)
• Owner: Frontend + Multimodal
• Priority: P2

lot-213-fx-rack [planned]

• Description: FX rack per track (chain multiple effects)
• Owner: Multimodal
• Priority: P2

lot-214-sample-browser [planned]

• Description: Sample browser (upload + reuse audio samples)
• Owner: Frontend + Backend API
• Priority: P2

lot-215-spectral-view [planned]

• Description: Spectral view (canvas visualization)
• Owner: Frontend
• Priority: P3

lot-226-stutter [done]

• Summary: /stutter glitch effect, /pan stereo, /master mastering chain

lot-230-compose-ui [done]

• Summary: ComposePage 3 button rows, per-track Solo/Loop/Reverse, status bar

lot-234-imagine-ui [done]

• Summary: ImaginePage upgrade, MediaExplorer compositions tab

lot-236-plan-docs [done]

• Summary: PLAN + docs update to 236 lots

lot-278-glitch [done]

• /glitch: random audio glitch effect (bit crush + tremolo + phaser)

lot-279-stretch [done]

• /stretch: time stretch without pitch change (atempo)

lot-280-final [done]

• Summary: 281 lots, 89 commands, DAW Phase 1-4, composition pipeline complete

lot-283-composition-browser [done]

• Description: Composition browser in Gestion tab (REST API fetch)
• Owner: Frontend
• Summary: Fetch compositions from /api/v2/media/compositions, display as buttons in Gestion tab, /comp load on click.

lot-284-track-type-badges [done]

• Description: Track type indicators (M/V/N badges) in timeline blocks
• Owner: Frontend
• Summary: typeIcon() badge in TimelineGrid blocks, CSS .daw-block-badge.

lot-285-e2e-health [done]

• Description: Full E2E + health check + PLAN update
• Owner: Ops
• Summary: Health 7/7, E2E verified, PLAN updated to 286 lots.

lot-286-build-deploy [done]

• Description: Build + deploy + commit
• Owner: Ops
• Summary: Vite build, Docker rebuild, git push.

lot-287-daw-ux [done]

• Summary: DAW UX overhaul — LCD display, faders, gradients, snap grid, loop region, animations

lot-288-backend-integration [done]

• Summary: REST endpoints for compositions/tracks, comp load sync, M/S/vol/pan wired

lot-290-e2e-daw [done]

• Summary: E2E DAW test verified, 310+ lots milestone

lot-300-opendaw [done]

• Summary: openDAW cloned, built, served on /daw/ with COOP/COEP, AI Bridge :8301

lot-305-compose-simple [done]

• Summary: ComposePage simple 4-track (MUSIQUE/VOIX/TEXTURE/EFFET), timeline, mix/master

lot-310-mini-daw-ia [done]

• Summary: Mini DAW IA natif -- 4 pistes, timeline drag/resize/dup, AI suggest, auto-compose, generate-all, 10 FX context menu, /api/v2/ai/suggest-prompt

lot-315-session-final [done]

• Summary: 315+ lots, 90 commands, 14 services, E2E 10/10, responsive verified

lot-531-context-perf [done]

• Summary: Context parallel load (Promise.all), command ranking (startsWith priority), mobile CSS deep pass, route prefetch (requestIdleCallback)

lot-532-mascarade-sse [done]

• Summary: mascarade SSE streaming confirmed working, thinking field extraction from Python side

lot-533-grafana-prometheus [done]

• Summary: Grafana + Prometheus monitoring stack, enriched /api/v2/metrics endpoint

lot-534-deepresearch [done]

• Description: /deepresearch multi-step search agent inspired by OpenSeeker
• Summary: deep-research.ts agent loop (plan→search→visit→extract→refine→synthesize), SSRF-safe page fetcher, LLM-driven query planning, /deepresearch and /dr commands with real-time progress

lot-535-thinking-fix [done]

• Description: mascarade thinking field extraction + streaming think suppression
• Summary: Python side (ollama.py): extract content from thinking field, strip <think> from streaming. Node.js side (llm-client.ts): strip inline <think>, suppress in SSE stream, clean final text. Double safety net.

lot-536-grafana-dashboard [done]

• Description: Grafana dashboard 16 panels + Prometheus multi-machine scrape
• Summary: Top stats (WS, chat/min, LLM calls, tokens, RSS, uptime), latency timeseries (HTTP, Ollama TTFB, persona, RAG), errors by label, throughput, counters table. Scrapes kxkm-ai + tower + mascarade.

lot-537-cloudflare-tunnels [done]

• Description: Cloudflare Tunnels for kxkm-ai + tower (replaces lots 538-540)
• Summary: 2 tunnels (kxkm-ai-gpu + intello-cockpit), 7 public hostnames, HTTPS auto. kxkm.saillant.cc (app), kxkm-api (API), kxkm-mascarade (LLM), kxkm-comfy (ComfyUI), kxkm-tower, kxkm-grafana, mascarade (Authentik SSO). Anthropic domain verification TXT added.

Phase Execution Immediate 2026-03-24 (P0/P1) [done]

Objectif

• Executer un lot de fiabilisation en 4 commits atomiques, testables et rollbackables.

lot-541-regex-intent-fix [done]

• Description: Corriger l extraction des prompts auto image/musique dans le routeur conversation.
• Owner: Backend API
• Fichiers cibles: apps/api/src/ws-conversation-router.ts, apps/api/src/ws-conversation-router.test.ts
• Changes: remplacer les motifs de regex \s* mal escapes, ajouter tests unitaires de non-regression.
• Checks: npm run -w @kxkm/api test
• Done criteria: prompts correctement extraits pour /imagine et /compose depuis langage naturel.

lot-542-audio-eventloop-unblock [done]

• Description: Sortir les traitements ffmpeg bloquants du chemin synchrone WS.
• Owner: Backend API + Multimodal
• Fichiers cibles: apps/api/src/ws-commands-generate.ts
• Changes: remplacer les appels execFileSync critiques par execution asynchrone via scheduler/subprocess non bloquant.
• Checks: npm run -w @kxkm/api test, smoke WS compose/mix
• Done criteria: pas de blocage event-loop sur mix/export, latence WS stable sous charge.

lot-543-worker-runtime-unification [done]

• Description: Unifier l execution worker autour de worker-runtime pour eviter la divergence index/runtime.
• Owner: Node Engine
• Fichiers cibles: apps/worker/src/index.ts, apps/worker/src/worker-runtime.ts, apps/worker/src/worker-runtime.test.ts
• Changes: index.ts devient orchestrateur fin, logique run/poll centralisee dans worker-runtime.
• Checks: npm run -w @kxkm/worker test
• Done criteria: une seule source de verite pour le cycle run/dequeue/execute.

lot-544-chat-render-budget [done]

• Description: Reduire le cout render de Chat.tsx pour les longues conversations.
• Owner: Frontend
• Fichiers cibles: apps/web/src/components/Chat.tsx
• Changes: memoization des derivees couteuses (compteur de mots, highlights), maintien UX inchange.
• Checks: npm run -w @kxkm/web test, npm run -w @kxkm/web build
• Done criteria: baisse du cout CPU render sans regression fonctionnelle.

Sequencement d execution

1. lot-541 (P0)
2. lot-542 (P1 fiabilite runtime)
3. lot-543 (P1 architecture worker)
4. lot-544 (P1 perf frontend)

Checkpoints

• J0: lot-541 merge + tests API verts
• J1: lot-542 merge + smoke compose/mix OK
• J2: lot-543 merge + tests worker verts
• J3: lot-544 merge + build web OK

Risques cibles

• R-INTENT-REGEX: extraction de prompt incorrecte
• R-AUDIO-BLOCKING: blocage event-loop API
• R-WORKER-DRIFT: divergence entre runtime et index
• R-CHAT-RENDER: degradation perf UI sur historique long

Phase Execution Next 2026-03-24 (P1/P2) [in-progress]

Objectif

• Enchainer directement sur le prochain lot de coherence plan/docs puis sur le pipeline composition prioritaire.

lot-545-plan-todo-sync [done]

• Description: Harmoniser PLAN.md, TODO.md et docs/AGENTS.md sur les statuts reels des lots 95-100 et 178-200.
• Owner: Coordinateur
• Fichiers cibles: PLAN.md, TODO.md, docs/AGENTS.md
• Checks: revue manuelle des statuts, coherence des priorites P1/P2
• Done criteria: plus de divergences de statut entre fichiers de pilotage.

lot-546-compose-api-direct [done]

• Description: Demarrer lot-178 avec appel ACE-Step direct pour fiabiliser la duree /compose.
• Owner: Backend API + Multimodal
• Fichiers cibles: apps/api/src/ws-commands-generate.ts, scripts/ ou service ACE-Step associe
• Checks: test manuel /compose 5s, 30s, 120s
• Done criteria: duree demandee respectee sans drift observable.
• Status: implementation validee, fallback en place, lot cloture.

lot-547-timeline-model-v1 [in-progress]

• Description: Demarrer lot-180 avec modele timeline v1 (tracks, clips, markers).
• Owner: Backend API
• Fichiers cibles: apps/api/src/composition, packages/domain si necessaire
• Checks: tests unitaires modele + validation schema
• Done criteria: structure timeline stable et serialisable.

lot-548-waveform-ui-v1 [planned]

• Description: Demarrer lot-194 avec waveform timeline cote frontend.
• Owner: Frontend
• Fichiers cibles: apps/web/src/components/compose ou timeline
• Checks: npm run -w @kxkm/web test, npm run -w @kxkm/web build
• Done criteria: waveform lisible par piste sans regression mobile.

lot-549-composition-tests [planned]

• Description: Renforcer les tests composition (store, remix, export, timeline path).
• Owner: Backend API + Frontend
• Fichiers cibles: apps/api/src//*.test.ts, apps/web/src//*.test.tsx
• Checks: npm run test:v2
• Done criteria: couverture non-regression sur le pipeline composition.

Sequencement d execution

1. lot-545 [done]
2. lot-546 [done]
3. lot-547 [in-progress]
4. lot-548
5. lot-549

Risques cibles

• R-DOC-DRIFT: statuts incoherents entre fichiers de pilotage
• R-COMPOSE-DURATION: duree audio non respectee en production
• R-TIMELINE-MODEL: format instable pour les futures features DAW
• R-WAVEFORM-UX: surcharge render sur mobile

Delta Session 2026-03-24

• lot-192-ws-commands-extraction est cloture.
• Phases 1-3 executees: extraction de /comp, /layer, /mix, /voice, /tracks, /undo, /solo, /unsolo, /rename, /dup, /bpm, /clear-comp, /preview, /gain, /loop, /swap, /info, /concat, /silence, /template, /marker, /metronome, /delete, /suggest, /snapshot et /randomize vers apps/api/src/ws-commands-compose.ts.
• Validation: diagnostics TS OK, suite @kxkm/api relancee avec ws-commands.test.ts OK, cloture documentaire lot-552 appliquee.
• Prochaine etape: lot-549 tests composition, puis lot-548 waveform/timeline UI.

Delta Session 2026-03-24 — phase 3

• lot-192-ws-commands-extraction: phases 1-3 executees.
• Bloc compose complet extrait vers apps/api/src/ws-commands-compose.ts, incluant maintenant concat/silence/template/marker/metronome/delete/suggest/snapshot/randomize.
• Validation: diagnostics TS OK, suite @kxkm/api relancee avec ws-commands.test.ts OK.
• lot-552: synchronisation PLAN/TODO/AGENTS + rapport QA webdesign + veille OSS ciblee sur waveform/tests/orchestration.
• Prochaine etape: lot-549 tests composition, puis lot-194 waveform/timeline UI.

Delta Session 2026-03-24 — validation compose + waveform

• lot-549-composition-tests est valide.
• Couverture ajoutee sur apps/api/src/composition-store.test.ts et apps/api/src/ws-commands.test.ts pour rebind composition active, clamps timeline, /comp load, /mix, /snapshot, /template et /delete.
• Validation lot-549: ../../node_modules/.bin/tsx --test src/composition-store.test.ts src/ws-commands.test.ts => 44/44 tests OK.
• lot-548-waveform-ui-v1 est implemente dans apps/web/src/components/ComposePage.tsx avec wavesurfer.js pour les apercus de pistes et blocs timeline.
• Validation lot-548: npm run check dans apps/web OK.
• Prochaine etape recommandee: capture visuelle Playwright du compose timeline + nettoyage du doublon ComposePage 2.tsx.

Delta Session 2026-03-24 — cleanup compose

• lot-553-cleanup-compose-doublon execute.
• Suppression du fichier duplique non reference apps/web/src/components/ComposePage 2.tsx.
• Validation: npm run check dans apps/web OK.
• Risque reduit: suppression d'une source de divergence entre UI compose legacy et UI active.

Delta Session 2026-03-24 — QA visuelle compose

• lot-554-qa-compose-timeline execute.
• Validation Playwright ciblee sur apps/web/e2e/visual-qa.spec.ts (grep ComposePage|composition timeline).
• Resultat: 2/2 tests OK, captures regenerees 10-compose-page.png et 18-composition-timeline.png.
• Conclusion: rendu Compose/timeline stable apres migration waveform + cleanup doublon.