11 projects
demystify-platform-cache
Production caching seam for AI workloads (Python mirror): exact-match cache, verified semantic cache, embedding cache, singleflight request coalescing, negative cache, and epoch-based invalidation. In-memory reference default; Redis extra.
demystify-voice
Interruption-aware voice agent runtime (barge-in, spoken-prefix truncation, multi-turn context repair)
demystify-voice-lingua
Low-latency multilingual extension for the Demystify voice pipeline: in-stream LID, hysteresis-damped mid-call language switching, per-language STT/TTS routing
demystify-rag
Client SDK for dmstfy-rag: grounded RAG queries with citations, groundedness, abstention and SSE streaming
demystify-extract
Client SDK for dmstfy-extract: schema-first document extraction with generated Pydantic types
demystify-events
Durable jobs/events for Demystify: asyncio mirror of @demystify/events (enqueue, transactional outbox, idempotent workers, backoff+jitter, DLQ).
demystify-gateway-sdk
Python SDK for the dmstfy-gateway data plane (chat, embeddings, STT, TTS via one OpenAI-compatible endpoint)
demystify-platform-state
Redis-backed shared-state seam with an in-memory reference (Python mirror): KV+TTL, atomic reserve/reconcile counter, priority queue, leased registry, and pub/sub. Backs gateway multi-instance correctness and the voice session registry.
demystify-platform-profiles
Unified Demystify deployment profiles (Python mirror): typed profile models, the three canonical profiles as data, env-overlay loader, and the canonical tenancy migration profile pair.
demystify-platform-observability
Shared metrics + dashboard for Demystify services (Python mirror): canonical RED metric names, a prometheus-client registry with RED helpers, a framework-neutral /metrics handler, an offline HTML dashboard, and a span->metrics bridge.
demystify-platform-contracts
Single source of truth for Demystify cross-cutting wire contracts (Python mirror): error taxonomy, domain events, canonical headers, key format/hashing, currency helpers, key-introspection models.
