Show HN: Agent-estimate, how long a coding task takes, at agent speed (github.com)

A new tool called agent-estimate has been introduced, enabling developers to quickly estimate the time required for coding tasks executed by AI agents. By utilizing three-point PERT (Program Evaluation and Review Technique) estimates, the tool provides not only the expected completion time but also a human-equivalent comparison and a compression ratio. This capability addresses a critical challenge in project management—accurately forecasting how long AI-driven tasks will take—helping teams avoid bottlenecks associated with scope creep and missed deadlines. The significance of agent-estimate lies in its ability to leverage multiple AI models, such as Opus 4.7, GPT-5.5, and Gemini 3.1, each with varying reliability horizons and costs. This feature allows for a precise calibration of estimates based on actual agent performance, therefore, offering a holistic view when determining task execution. Users can derive insights rapidly—installing the plugin takes only 30 seconds, after which estimates can be generated almost instantaneously, allowing developers to allocate resources effectively across tasks. Overall, agent-estimate streamlines project planning and execution, enhancing the efficacy of AI/ML workflows in software development.