Fluxor Documentation
Fluxor (formerly Heliox) is a visual-first, local-first, Apache-2.0 IDE for designing agentic flows — DAGs of LLM and tool-calling steps — and integrating them into external systems through portable, conformance-verified cross-runtime SDKs. It’s bring-your-own-key: your provider tokens and your flows never leave your machine unless you choose to serve or export them.
These guides take you from install to serving flows in production.
Start here
- Getting Started — Install Fluxor, design your first agentic flow, run it, and export a portable
.flow.json. - Core Concepts — The mental model behind Fluxor: steps, flows, loops, checkpoints, roles, mods, and contracts.
Build flows
- Flow Design — Designing agentic flows on the canvas or with Text-to-Flow, attaching roles and mods, and running at any granularity.
- Roles & Mods — The signed marketplace catalog — 12 flows, 14 roles, 33 mods, and 7 steps — and how they attach and compose.
- Provider Connections — Bring your own key: DBeaver-style connection profiles for any OpenAI- or Anthropic-compatible endpoint.
Integrate & ship
- Cross-Runtime SDKs — Design a flow once, export a portable
.flow.json, and run it identically on the TypeScript, Java, and Python runtimes. - Serve & Deploy — Expose an exported flow over HTTP or as an MCP server, and trigger it on a webhook or a schedule.
- API Reference — The
.flow.jsonexport schema, SDK entry points per runtime, and the serve HTTP/MCP surface.
Reference
- Architecture — The real layering behind Fluxor: from the canvas down through execution, export, cross-runtime SDKs, serving, and the signed market.
- Contributing — Where the code lives, the one rule that’s non-negotiable, and how to keep a PR reviewable.