Mastra introduces durable agent goals

// 2h agoPRODUCT UPDATE

Mastra introduces durable agent goals

ANNOUNCEMENT PRODUCT GITHUB PRODUCT HUNT X

The Mastra 1.42.0 update introduces an experimental Goals feature that allows TypeScript-based AI agents to pursue durable, thread-scoped objectives. Supported by a database backend, agents use an in-loop LLM-as-a-judge mechanism to evaluate progress across process restarts and multi-step conversations until the objective is met.

// ANALYSIS

Evaluating agentic behaviors using LLM-as-a-judge directly in the loop is a major step forward for building reliable, autonomous agents in TypeScript.

* Persistent Objectives: Moving away from ephemeral prompt instructions to database-backed, thread-scoped goals ensures agents stay on track even if the application environment restarts or gets interrupted.

* In-Loop LLM Judging: Automating success criteria check-ins directly within the agent loop reduces manual boilerplate and enables dynamic, self-terminating agent sessions.

* TypeScript Ecosystem Boost: As an enterprise-focused TypeScript framework, Mastra is positioning itself as a serious competitor to Python's LangChain and CrewAI for building robust production agents.

* Experimental Caveat: Since the API is marked as experimental, developers should expect potential breaking changes in storage schemas or configurations in future minor releases.

// TAGS

mastratypescriptagentllm-as-a-judgeopen-sourcedevtool

DISCOVERED

2h ago

2026-06-15

PUBLISHED

2h ago

2026-06-15

RELEVANCE

8/ 10

AUTHOR

mastra

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

LAUNCH34m ago

Inception's Mercury 2, the first commercial-scale reasoning diffusion LLM, is now available for production deployment on Baseten.

Baseten has announced that Inception's Mercury 2 is now live on its platform, making it the first inference platform to deliver production-grade reasoning diffusion LLMs (dLLMs) to developers. Unlike traditional autoregressive models that generate tokens sequentially, Mercury 2 uses a diffusion architecture to generate and refine multiple tokens in parallel, enabling speeds of over 1,000 tokens per second on widely-deployed NVIDIA GPUs. Partners like Augment Code have already deployed Mercury 2 in production, achieving a 90% reduction in inference costs and an 82% drop in latency for critical workloads, while maintaining quality comparable to speed-optimized models like Claude 3 Haiku and GPT-5 mini.

OPEN SOURCE36m ago

optimizerDuck is a free, open-source C# utility for Windows designed to boost performance, enhance user privacy, and clean up system bloat.

optimizerDuck is an open-source Windows optimization utility written in C# that helps users fine-tune their operating systems. The tool features a clean interface that allows users to easily toggle Windows telemetry settings, manage startup applications, uninstall bloatware, and optimize system configuration for better performance. By providing a transparent, community-driven alternative to proprietary PC cleaners, it addresses user concerns regarding privacy and bloatware on modern Windows installations.

OPEN SOURCE43m ago

TeslaMate is a self-hosted, open-source data logger for Tesla vehicles that captures detailed telemetry, trip history, and charging metrics while keeping user data private.

TeslaMate is a self-hosted data logging application designed for Tesla owners who want to track their vehicle's telemetry without relying on third-party cloud services. Written in Elixir and leveraging Grafana for visualization, TeslaMate securely collects data from the Tesla API, recording trip logs, charging sessions, battery degradation, sleep cycles, and firmware updates. By hosting it locally or on a personal server (typically via Docker), users maintain complete ownership of their driving data while gaining access to highly detailed dashboards and statistics.