Rheosoph/flow-like

What it solves

Flow-Like is a unified platform that eliminates the need to stitch together separate tools for automation, data storage, AI integration, and app building. It provides a single, local-first runtime where users can build end-to-end use cases—such as private RAG assistants or large-scale data pipelines—without relying on external cloud services or complex glue code.

How it works

The platform uses a visual workflow studio (a no-code IDE) to build typed, traceable execution paths. Its core engine is written in Rust for high performance, allowing it to run on devices ranging from laptops and servers to phones and edge devices. It features a native data layer that supports files, tables, and hybrid keyword/vector search, enabling it to hold terabytes of data in-platform. AI capabilities, including LLMs, embeddings, and speech-to-text, are integrated as nodes that can run locally on-device or in the cloud.

Who it’s for

It is designed for developers, data analysts, and business users who want to complete autonomy over their data and infrastructure, avoiding vendor lock-in and cloud costs. It is particularly useful for those operating in air-gapped or secure environments where data privacy is paramount.

Highlights

• Local-first and Offline-capable: Runs entirely on your own hardware (laptop, server, Raspberry Pi) with no forced cloud dependency.
• Native Data Layer: Built-in storage for terabytes of data and hybrid search, removing the need for external databases like S3 or Postgres.
• High Performance: Rust-based engine capable of processing hundreds of thousands of workflows per second with sub-millisecond latency.
• AI-native: Local execution of LLMs, vision, and speech models with full reasoning traces and guardrails.
• Extensible: Custom nodes can be written in 15 different languages via sandboxed WebAssembly.
• App Shipping: Ability to package workflows as installable apps with a no-code UI builder.