05·The Project
Core Engine Build Status
Epic-level progress tracking for the GamiWays Core Engine, manually synced from the development repository.
Synced from repository2026-08-02
gami-lab/gamiways-corePhase B underway
The Phase A baseline is delivered: conversational runtime, async Game Master, 3-layer memory, Context Engine v2, typed RAG, Scenario Builder and public Web App — 683 commits in 105 days. Phase B now extends the experience with voice, media and a richer player frontend.Current programme phasePhase B underway
34 Delivered baseline
1 Phase B
5 Phase B milestones
Per-phase breakdown
Phase A — Minimal CoreApril → July 2026
34/34 · 100%Phase B — Enhanced ExperiencesAugust → December 2026
UnderwayPhase C — Research & Scale ReadinessTBD
0/3 · 0%PHASE AApril → July 2026
Phase A — Minimal Core
Build and validate the fundamental loop: user input → context assembly → orchestrated avatar response → memory update. Deliver operational back-office, a public text-based player app, and one real scenario prototype.
34/34
epics
1.1Platform SetupDone
Monorepo (pnpm + Turborepo), TypeScript strict, Docker Compose, CI baseline. Completed 2026-04-22.
1.2First LLM Loop + ObservabilityDone
Provider abstraction layer, OpenAI/Anthropic/Mistral adapters, Langfuse integration, /v1/exchange, API-key auth, latency/token metrics. Completed 2026-04-22.
2.1Avatar Agent v1Done
Avatar domain model, persona prompt assembly, multi-turn conversation flow, message persistence, POST /v1/conversations/{id}/messages. Completed 2026-04-22.
2.1bAvatar Agent v2 — Memory + RAGDone
Context-aware avatars: user persona injection, working memory, long-term facts, typed RAG retrieval (memory|world|media). Canonical shared DTO ownership. Completed 2026-05-13.
2.2Scenario & Session LifecycleDone
Scenarios, avatars, sessions, conversations, lifecycle APIs, explicit conversation model, session reset, avatar CRUD, scenario CRUD, compaction triggers. Completed 2026-04-22.
2.5Admin CRUD + Console IntegrationDone
Admin CRUD APIs, session admin tooling, inline scenario/avatar editing, operational reset tooling, console integration, conflict-safe deletes. Completed 2026-04-28.
2.6GM Debug Panel v1Done
GM event inspection, transition history, unlocked avatar inspection, GM notes, admin session inspection, safe observability APIs, event filtering. Completed 2026-04-28.
2.7Runtime Inspector v2Done
Unified runtime inspector, assembled context inspection, runtime metrics, memory inspection, SSE event integration, runtime admin actions (GM replay, memory refresh, memory clear). Completed 2026-05-07.
2.8Console Debugging RedesignDone
Unified debugging shell, memory evolution workspace, GM causality trace, turn profiler, persona-first flow, consolidated debug navigation, runtime visualization redesign. Completed 2026-05-07.
3.1Health & Dependency MonitoringDone
Dependency probes, admin health endpoint, degraded-state reporting, Redis/Postgres/LLM checks, runtime dependency inspection. Completed 2026-04-30.
3.2Inspector Consolidation & Contract CleanupDone
Inspector contract ownership cleanup, DTO canonicalization, backend route consolidation, console flow unification, duplicate DTO removal, inspector hardening tests. Completed 2026-05-10.
4.1Async Game Master v1Done
Async GM orchestration, avatar unlock logic, routing decisions, GM directives, event-driven orchestration, AI Guided Discovery reference scenario, runtime GM observability. Completed 2026-04-29.
4.1cMulti-Model Runtime ConfigurationDone
Role-based model resolution, global/role/avatar overrides, OpenAI/Anthropic/Mistral/xAI adapter registry, admin model-config endpoints, console editor, and effective model diagnostics. Completed 2026-05-20.
4.2Memory Layer v1Done
User memory facts, fact extraction, long-term user memory, memory injection into prompts, memory admin APIs. Completed 2026-05-05.
4.2bMemory System v2Done
Layered memory contracts, working memory storage, async maintenance pipeline, working-memory repositories, avatar memory context assembly, memory inspection APIs, deterministic layered memory injection. Completed 2026-05-06.
4.2cEpisodic + Hydrated Memory SystemDone
Episodic persistence, hydration pipeline, deterministic memory selection, GM memory integration, memory observability, conversation working memory, hydration diagnostics, memory selection policies. Completed 2026-05-08.
4.3Performance BaselineDone
Turn metrics, GM metrics, latency tracking, token tracking, metrics aggregation, admin metrics endpoint, event-log timing persistence. Completed 2026-04-30.
4.4Multi-Avatar Navigation v1Done
Unlockable avatars, avatar switching, session-scoped unlock progression, avatar awareness, AI Guided Discovery scenario, availability APIs. Completed 2026-04-27.
4.5Runtime State & SSE EventsDone
Runtime state endpoint, SSE event streaming, async runtime events, processing state tracking, runtime pub/sub layer. Completed 2026-05-05.
5.1Knowledge Substrate — Ingestion & RetrievalDone
Canonical knowledge/retrieval contracts, persistence schema, typed ingestion lifecycle (queued→running→completed|failed), type-specific retrieval (memory|world|media), knowledge API endpoints, stack-e2e coverage. Completed 2026-05-11.
5.1bAvatar-Scoped Knowledge VisibilityDone
Knowledge sources and chunks now support visibleToAvatarIds, avatar retrieval is visibility-filtered, GM retrieval remains unrestricted for orchestration, and bounded diagnostics expose visibility decisions. Completed 2026-05-21.
5.2Context Engine v2Done
Dedicated Context Engine module with explicit input/output/trace contracts, deterministic precedence policy, token-budget selection telemetry, machine-readable selection/trimming trace, conflict resolution for long-term facts and retrieval chunks. Completed 2026-05-11.
5.5User Persona SystemDone
User personas, persona persistence, avatar persona injection, GM persona injection, persona APIs (GET/PUT /v1/users/{userId}/persona). Completed 2026-05-02.
7.1Public User Web App v1Done
Production-facing player app in apps/web: local identity + persona onboarding, scenario selection, available-avatars-only discovery with SSE unlock updates, single active chat with optimistic send and explicit conversation end. Coolify-ready static deploy (VITE_API_URL + VITE_API_KEY). Completed 2026-06-01.
5.3UX Streaming LayerDone
Public message-stream client and progressive avatar rendering in apps/web. Contiguous delta ordering, exact-once avatar persistence, completion/interruption cleanup, provider iterator and web reader cleanup. Additive SSE transport: POST /v1/conversations/{id}/messages/stream. Completed 2026-07-29.
6.1Scenario Builder v1Done
Scenario/avatar editors, knowledge-source authoring, visibility policy, model selection per scenario/avatar, admin audit remediation. Completed 2026-07-12/18.
8.1Avatar Trait StructuringDone
Explicit scenario-scoped avatar-trait preparation, canonical computedTraits schema and persistence, admin trigger + read-only inspection endpoint. Completed 2026-07-19.
8.2Runtime Context Assembly RefactoringDone
Structured runtime context sections, trait-aware avatar prompt assembly wired through send-message, runtime context snapshots aligned in inspector. Completed 2026-07-20.
8.3Game Master Prompt RefinementDone
Explicit GM prompt structure with prompt contract ownership, stronger decision-policy guidance, hardened test coverage. Completed 2026-07-20.
8.4Working Memory Prompt RefinementDone
Structured working memory with coveredTopics and unresolvedThreads fields, refined compaction normalization, aligned operator surfaces. Phase A complete. Completed 2026-07-20.
8.5Game Master Post-Analysis RefinementDone
Explicit dialogue-control modes (user_led, avatar_guided, avatar_led, repair, transition), first-class retrieval planning for next Avatar turn, simplified routing contract, dynamic routing capabilities, removal of application-owned fields from GM output, required non-empty Director Notes. Completed 2026-07-29.
8.6Scripted Conversation EvaluationDone
TypeScript CLI evaluation tool: versioned JSON definitions, sequential session execution, semantic LLM judge via /v1/exchange, per-question latency/token/cost metrics, atomic partial-result persistence, multi-model comparison reports, local dependency-free report viewer. Completed 2026-08-02.
6.2Real Scenario ValidationDone
AI Guided Discovery scenario validated end-to-end. Scripted evaluation tool (EPIC 8.6) enables repeatable quality and latency benchmarking across model configurations. Completed 2026-08-02.
6.3Summer Prototype DeliveryDone
Phase A complete: headless conversational runtime, operational admin back-office (apps/admin), public player web app (apps/web), AI Guided Discovery reference scenario, scripted evaluation tooling. Demoable to external stakeholders. Completed 2026-08-02.
PHASE BAugust → December 2026
Phase B — Enhanced Experiences
Phase B underway: extend the delivered Core with voice input/output, multimedia triggers, multiple scenarios, richer memory systems, and a user-facing frontend.
0/5
epics
B.0Phase B ActivationIn Progress
Transition from the delivered Core baseline to the enriched-experience workstream, with voice, media and player experience priorities.
5.4Guided Progression EnginePlanned
Scenario objectives tracking, progression state machine, GM-driven narrative milestones
B.1Voice Integration (STT + TTS)Planned
STT pipeline (Deepgram / Whisper), TTS streaming (Cartesia / Inworld TTS-2), real-time audio delivery
B.2Multimedia TriggersPlanned
GM-driven media events: image display, video playback, document reveal during conversation
B.3User-Facing Frontend v2Planned
Richer player experience beyond EPIC 7.1: session history, progression visualization, multi-device support, voice-ready UI
PHASE CTBD
Phase C — Research & Scale Readiness
Prepare the platform for advanced integrations: expressive avatars, advanced persona systems, scaling, SDKs, and research partnerships.
0/3
epics
C.1Expressive Avatar IntegrationPlanned
Real-time video avatar binding, body language triggers, lip-sync coordination
C.2Multi-tenancy & SecurityPlanned
Tenant isolation, JWT auth, RBAC, data residency controls
C.3SDK & API ProductizationPlanned
Public SDK, versioned API, developer documentation, integration examples
Editorial note — This page is manually synced from the development repository gami-lab/gamiways-core . It reflects the actual project state as of the portal's last update date.