Skip to content

1.0.0

Latest
Latest

Choose a tag to compare

View all tags
@Vensus137 Vensus137 released this 28 Feb 12:10
· 1 commit to main since this release
1.0.0
f2d92f0

[1.0.0] - 2026-02-28

First stable release of Coreness Flow.

Coreness Flow is an event-driven AI assistant that runs entirely on your machine. Unlike chat-only tools, it reacts to events — messages, webhooks, schedules, plugin signals — and runs YAML-defined scenarios: chains of steps with LLM calls, branching, and integrations. Everything stays local, under your control.

This release delivers the full stack: desktop UI, scenario engine, plugin system with UI contributions, and local RAG — ready for automation and extension.

Highlights

  • Event-driven scenarios — Triggers (chat message, webhook, cron), step chains, conditional branching, and cross-scenario calls. Logic lives in YAML; no code required for typical flows. Data moves through _cache and placeholders.
  • Plugin system & UI contributions — VS Code–style plugins: config.json defines metadata, settings, actions, and contributes (tabs, sidebar items, settings). Frontend builds the UI from contributions; new integrations are new plugin folders.
  • Async actions — Long-running work runs in the background via call_nowait; scenarios continue without blocking. Completion is checked via _async_action placeholders.
  • LLM & local RAG — Request routing by complexity (cheaper models for simple tasks, stronger for complex). Local RAG: BGE-M3 ONNX INT8 + embedded Qdrant; no mandatory external vector services.
  • Desktop app — Electron + React UI, Python 3.11 backend, WebSocket and API Bus (actions + events). SQLite and JSON for config; YAML for scenarios. Windows 10+.

Added

  • YAML scenario engine with triggers (message, webhook, cron), steps, transitions, and cross-scenario invocation
  • Plugin discovery via config.json; registration of actions and UI contributions (tabs, sidebar, settings)
  • Async action execution via call_nowait and _async_action placeholders for non-blocking flows
  • LLM routing by complexity; local RAG pipeline (BGE-M3 ONNX INT8, Qdrant embedded)
  • Full documentation: architecture, plugins, scenarios, actions, events, placeholders, UI contributions

Technical

  • Layered architecture: UI (Electron + React) and backend (Python plugins) communicate only via API Bus (WebSocket)
  • Single-user, local-first; config and data under user control (e.g. %APPDATA%\CorenessFlow)
Assets 3
Loading