SShortSingh.
Back to feed

Developer Builds Custom AI Coding Agent 'Dasan' Using OpenAI Codex OAuth

0
·1 views

A developer has created a personal AI agent called Dasan, named after the pen name of Joseon-era scholar Jeong Yak-yong, rather than relying on existing tools like Claude Code. The agent runs on an LLM powered by Codex OAuth, a choice driven by the cost constraints of using paid API access freely. Dasan operates on a ReAct loop, enabling it to explore, read, modify, and execute commands on local files within a single persistent session stored via SQLite. The system uses a two-layer prompt architecture separating a fixed core role from a learnable user-alignment layer. The project is open source on GitHub and can be installed globally on macOS, Linux, or Windows with a single command.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingHacker News ·

PostgreSQL May Replace Multiple Specialized Systems in Your Stack

A discussion on Hacker News highlights a growing perspective that PostgreSQL alone can handle many infrastructure needs typically split across multiple systems. The argument centers on Postgres's broad feature set, which includes support for queuing, search, caching, and more. Developers are questioning whether adding separate specialized tools is always justified when Postgres already covers those use cases. The conversation is linked to a dedicated site, postgresisenough.dev, making the case for consolidating around a single database. The post has attracted community engagement on Hacker News, sparking debate about simplicity versus specialization in modern software stacks.

0
ProgrammingDEV Community ·

MongoDB vs PostgreSQL: A Practical Guide to Picking the Right Database in 2026

PostgreSQL and MongoDB are both mature, production-ready databases suited to different use cases rather than competing on quality. PostgreSQL is a relational database that enforces strict schemas, supports complex joins, and offers strong ACID transaction guarantees, making it ideal for financial systems, e-commerce, and SaaS platforms. MongoDB is a document-based NoSQL database that stores flexible JSON-like records, making it a natural fit for rapidly evolving data models, content management, and high-throughput applications. The key deciding factor is the shape of your data: highly relational, structured data favors PostgreSQL, while fluid, document-shaped, or object-driven data favors MongoDB. Developers are advised to treat the choice as a practical one based on data structure and access patterns rather than a high-stakes debate.