Insights, updates, and best practices from the Dosu AI teammate for developers. Stay up to date with the latest features, tips, and industry trends.
People have dreamed of living documentation–documentation that evolves as fast as your code–for decades but human bandwidth to review and update documentation has always been the bottleneck. Now with AI, we have a way to couple written knowledge to code and get the best of both worlds: easy to understand documentation that evolves with your code.
June Dosu Drop: One-click documentation, @ mentions, improved code reasoning, and the new fact-based reasoning agent now GA. Building the future of automated knowledge management.
Dosu's new fact-based reasoning agent is now generally available, featuring smarter problem-solving, customizable behavior controls, and improved knowledge management for development teams.
Learn how to implement distributed dependency injection for testing AI agents across multiple services, using Celery preserializers to handle complex dependencies in asynchronous workflows.
Introducing the official Dosu blog - your source for insights, updates, and best practices from the AI teammate that helps developers maintain their code.
Learn how to seamlessly integrate Pydantic models with Celery tasks using custom preserializers, eliminating manual serialization and type annotation issues.
Discover why comprehensive documentation is now essential for AI-powered development teams and how it can transform your engineering productivity in the age of intelligent automation.
May 2025 updates: New fact-based reasoning agent, self-serve Teams pricing plan, and refreshed website design with improved navigation and animations.
Let’s get this out of the way: no one gets into software because they love writing documentation.
April 2025 updates: Introducing Dosu Does Docs beta - AI-powered documentation generation that transforms features into shareable docs in minutes, plus community insights on automating GitHub issue triage.
Open-source software is built on the backs of maintainers, they're the ones who keep everything running smoothly. But lately, it feels like everyone's on the verge of burnout. Surveys show that almost 60% of maintainers have either quit their projects or thought about it 1. These aren't just numbers; we're talking about real people who are overwhelmed and stretched too thin.
We’ve all seen it, managing a constant influx of GitHub issues in a popular GitHub repository can quickly become overwhelming. Development teams and open-source maintainers alike face the daily challenge of sorting through new issues, determining their nature, and routing them to the right people. This administrative burden often takes valuable time away from actual coding and can significantly slow down project progress.
March 2025 updates: 40k+ installations milestone, self-serve Slack & Confluence integrations, and early access to Dosu Knows - our AI-powered knowledge base.
February 2025 updates: Introducing Spaces - share data sources across channels, plus new preview generation and inbox features for better team collaboration with Dosu.
A proper labeling scheme is one of the best ways to keep an open source project in GitHub organized. Labels are a helpful visual indicator, improve discoverability, and allow for automated analytics and reporting.
There are many techniques for “teaching” LLMs to improve their performance on specific tasks. The most common are:
Learn how Dosu uses evaluation-driven development and LangSmith to build reliable LLM products at scale, ensuring quality while iterating rapidly.
How Dosu uses Astronomer to power reliable data pipelines for AI applications, from ingestion to vector indexing and analytics.
