Skip to main content

SF Veritas Desktop App

AI made writing code fast, but trusting what ships became the hard part. That's why we built SF Veritas (SF = Sailfish)!

SF Veritas is a unified development platform designed to help developers ship faster with confidence. It aims to make AI-driven development more reliable and less fragile.

What SF Veritas gives you​

  • Agent independence — Works with any AI coding agent (Claude Code, Codex, Cursor, GitHub Copilot, and more). No lock-in; swap agents any time.
  • Multi-agent workflows, no babysitting — Orchestrate tasks across multiple agents and terminals. Queue work, track progress, and let agents run autonomously while you focus elsewhere.
  • Spec-driven development, done differently — Write lightweight specs with acceptance criteria, link them to code symbols, and let the system track what's implemented and what's drifted.
  • Automated verification layer — After an agent finishes a task, a separate validator agent checks the work against the spec. Failures loop back automatically—no manual QA pass needed.
  • AI debugging without breakpoints — Full-stack observability (Console, Flamechart, distributed traces) captured automatically via the SDK. Diagnose issues by inspecting real telemetry, not by adding console.log statements.

Architecture​

The SF Veritas Desktop App consists of three core components:

  1. Desktop App: The user interface that displays Console logs and Flamechart visualizations
  2. Local Collector Service: A lightweight Node.js backend that receives telemetry data from your instrumented applications
  3. Local SQLite Database: Stores telemetry data locally with configurable retention (default: 7 days, 1GB limit)

Data Collection Policy​

Sailfish does not collect information about your codebases, code, or product usage data.

The app collects only anonymous usage metrics to improve the product:

  • Feature usage timestamps
  • Command executions
  • Performance data (memory, render times)
  • Database statistics
  • System metadata (OS, IDE version)

All telemetry data from your applications stays on your local machine.

Getting Started​

  1. Install - Get the Desktop App installed
  2. Integrate with Your Code - Instrument your backend and frontend code
  3. Learn how to use it - Master the Console and Flamechart features

Requirements​

  • Node.js 16+ for instrumented applications
  • macOS, Windows, or Linux