Skip to content

o-henry/rail

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1,321 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RAIL

RAIL is a local-first Tauri desktop app for graph workflows, task threads, research collection, and workspace knowledge management.

It combines:

  • a DAG canvas for structured multi-step execution
  • a task thread surface for agent-driven work
  • a research pipeline that can collect, normalize, and visualize evidence
  • a local knowledge/database layer for stored runs, artifacts, and documents

RAIL is designed for fast iteration inside a single workspace without depending on a hosted orchestration backend.

What RAIL Does

RAIL supports three main working styles:

  1. Graph workflows Build node-based flows that combine role nodes, transforms, gates, and data collection.
  2. Task threads Ask for work in a chat-like thread, tag role agents such as @researcher, and review artifacts as they are produced.
  3. Research monitoring Run research-oriented collection jobs, inspect structured evidence, and view question-aware charts in the Visualize tab.

Core Surfaces

Graph

The Graph tab is the canvas-first workflow editor.

Typical use cases:

  • build multi-node pipelines
  • branch on pass / fail decisions
  • combine role outputs into a final document
  • inject files and grounded evidence into later nodes

The graph runtime is oriented around DAG execution and explicit node-to-node handoff.

Tasks

The Tasks tab is the fastest way to ask for work.

Key behavior:

  • create a thread
  • tag one or more role agents such as @researcher
  • stream status, logs, and artifacts into the thread
  • stop a running request from the composer
  • inspect related files and generated outputs in context

Tasks are useful when you want the system to choose the execution path for you instead of hand-building a graph.

Visualize

The Visualize tab is the research monitor.

It is intended for questions such as:

  • “What are the best-rated genres on Steam right now?”
  • “Compare community sentiment for these games.”
  • “Show the strongest evidence behind this research report.”

Visualize reads normalized research outputs and renders:

  • question-aware charts
  • timeline or aggregate tables when appropriate
  • evidence streams
  • research history / prior sessions

Database

The Database tab is the local knowledge browser.

Use it to:

  • inspect stored run artifacts
  • open grouped documents
  • review previously generated outputs
  • manage saved research and knowledge entries

Settings

The Settings tab contains operational controls for the app.

Current settings areas include:

  • appearance and base preferences
  • Web Connect / bridge status
  • account and Codex-related controls
  • memory and retention management
  • locale selection

Research Pipeline

RAIL includes a research-oriented collection path used by @researcher.

That flow can:

  • interpret the user question
  • choose a collection mode such as genre ranking or comparison
  • collect relevant evidence
  • normalize collected items into local storage
  • generate a report spec for Visualize

The current system is built so that new research runs can be viewed later instead of being lost after a single answer.

Web Connect

RAIL can expose a local bridge for browser-connected flows.

The bridge surface shows:

  • local bridge URL
  • full connection code for the extension or external client
  • restart and refresh controls

The bridge is intended for local workflow integration, not public deployment.

Storage Model

RAIL is local-first.

Important data is kept in the workspace, including:

  • task runs
  • studio role runs
  • collected research artifacts
  • normalized research storage
  • knowledge/database entries

This makes it possible to inspect or reuse prior work without depending on a remote service.

Internationalization

The current user-facing locale selector supports:

  • Korean
  • English

Some additional locale assets may still exist internally, but the current settings surface is intentionally limited to the actively supported options.

Tech Stack

  • Tauri
  • React
  • TypeScript
  • Vite

Development

Install dependencies:

npm install

Run the app in development:

npm run tauri:dev

Type-check:

./node_modules/.bin/tsc --noEmit

Current Focus

RAIL is actively evolving around:

  • reliable role-agent orchestration
  • research collection quality
  • better visualize/report generation
  • local knowledge and artifact management
  • low-noise desktop UX for everyday use

About

No description or website provided.

Topics

Resources

Security policy

Stars

Watchers

Forks

Packages

 
 
 

Contributors