What’s new.
The things we’ve shipped, most recent first. We list what’s actually live — nothing more.
Nexith Chat
Chat with Nexith Core in the browser — long context, streaming responses, and memory that carries across sessions.
Privacy egress gate
Every response — streaming, non-streaming, or served from cache — now passes a single allowlist gate on the way out.
Session memory
Pinned facts and auto-captured context, scoped per user and per app, retrieved at request time.
Semantic cache
Requests close to one already answered are replayed from cache in milliseconds — cached input billed at up to 90% off on a hit.
OpenAI-compatible API
Point your existing SDK at our endpoint. Chat completions and embeddings, with SSE streaming on every endpoint.
Usage analytics & API keys
Dashboards for token usage and cost, plus fine-grained keys with scoped permissions and per-key rate limits.