SouthForge AI is an independent software studio founded by Tyler South that builds focused AI-assisted software products. With a foundation built on over 15 years of experience in IT and security, the studio creates tools designed to be secure by default, reliable, and genuinely useful for real-world workflows rather than focusing on hype or bloat. The studio builds its products in the open, emphasizing transparent engineering, validation before release, and platform foundations that ensure consistency across web, mobile, and desktop environments.

SouthForge AI operates as a lean studio, utilizing a shared platform for account management, subscriptions, and cross-platform delivery to support its various applications. Each product is engineered to solve a specific, practical problem, ranging from AI infrastructure tools for developers to specialized apps for gig workers, truck drivers, and agencies. The studio maintains a commitment to open-source where applicable and promotes a culture of building in public, providing detailed engineering case studies that document the architecture, observability, and data pipelines powering its software.

Some of the key features are:

• Conduit: An open-source local MCP gateway that centralizes all MCP server configurations and authentication while reducing token overhead by approximately 90%.
• Burnwatch: A spend monitoring tool for autonomous AI agents that uses observe-only SDK ingestion to alert on anomalous payment activity without interfering with payment paths.
• LimitWatch: A tracking tool that provides historical data, pricing drift, and usage metrics for AI subscriptions by analyzing snapshots and real token burn.
• RigCast: A road-risk and weather forecasting tool for OTR truck drivers that provides trailer-aware risk assessments like crosswind susceptibility and chain-law alerts.
• ReadyGig: A recordkeeping application tailored for gig workers to track mileage, receipts, and earnings with native mobile support.
• ProjectForge: A proprietary validation harness that orchestrates profile-driven AI tasks with release gates and automated visual regression checks.

The studio operates by building products that are actively used in production environments, avoiding demos or waitlists as products. Development often begins with identifying a specific need or workflow bottleneck—such as the repeated configuration of MCP servers or the lack of truck-specific weather risk reporting—and then engineering a stable, focused solution. These tools are then integrated into the studio's shared platform, ensuring that support, privacy, and billing features remain consistent. The studio's engineering process involves rigorous testing, such as using Playwright for automated checks and patchright for scraping, to ensure reliability across all releases.

Some common use cases include:

• Streamlining AI Agent Development: Developers use Conduit to manage multiple MCP servers centrally and reduce token usage, while using Burnwatch to monitor and secure autonomous agent payment wallets against spikes or errors.
• Enterprise AI Budgeting: Organizations and power users use LimitWatch to compare plan quotas across different AI providers, helping to optimize subscription costs based on actual measured usage rather than advertised limits.
• Commercial Fleet Safety: Truck drivers and small fleets use RigCast to obtain precise, trailer-specific road risk assessments and weather alerts that factor in vehicle-specific physical vulnerabilities to wind and terrain.
• Gig Economy Management: Freelance workers utilize ReadyGig to securely capture and maintain records of their daily income, mileage, and business expenses for tax and reporting purposes.
• Agency Reporting: Agencies use the Rundown application to automate the generation of client-facing progress reports by pulling metrics directly from CRM data to create consistent, brand-aligned monthly narratives.