Szept is a macOS app that enables voice dictation and screen sharing with AI coding agents like Claude Code and Cursor. Users hold fn to dictate text, use Command-Shift-A to capture and annotate screenshots, and can record video with their microphone. The app includes source code access and uses Soniox for cloud-based speech recognition at minimal cost.
Developers are frustrated with rigid Plan Mode UIs in coding assistants, but the real solution is embedding planning as an internal runtime primitive rather than a user-facing feature. When plans are baked into the agent harness as structured state machines with checkpoints and acceptance criteria, they enable dynamic refinement and keep both human and AI aligned on concrete goals without slowing down workflow.
Compaction.dev is a tool that reduces input and output tokens for Claude Code, Codex, and Cursor by compressing model-visible content locally before sending to providers and shaping output during generation. It operates within existing editors without requiring new tools, achieving ~14% input reduction and ~25% output reduction in demonstrated runs, with free and paid account options for different compression features.
Cursor released two AI bots for software development: Rollouts monitors code changes from pull request to production, detects regressions, and can automatically revert problematic changes; Security Reviewer scans code for vulnerabilities and proposes fixes. Both tools automate repetitive post-PR tasks to accelerate safe code deployment.
A team shares prompt engineering tips for improving token efficiency in an LLM agent harness, based on lessons learned at Cursor.
Cursor has launched Projects, a feature enabling AI agents to manage large bodies of work like features, migrations, and full apps while maintaining context over months. Projects use a coordinator agent that delegates to thousands of subagents, runs on cloud infrastructure, and can operate autonomously through subscriptions to signals like Slack channels and PR activity. Early metrics show new users merge 30% more PRs, while primary Project users merge six times as many.
Vigil is a macOS utility that keeps your computer awake while coding agents like Claude Code, Cursor, and Copilot CLI are working. It uses lifecycle hooks to detect agent activity and automatically releases the hold when tasks complete, fail, or timeout, with optional display control and menu bar monitoring. Priced at $3.99 one-time for two Macs, it prioritizes privacy by discarding conversation data and limiting network requests to updates and licensing only.
Owl24.dev is a production error monitoring platform that automatically detects and groups failures, then queues them for coding agents (Claude Code, Cursor) to investigate and submit pull requests for human review. The service handles error detection, triage, root-cause analysis, and fix generation while keeping humans in control of what reaches production.
Orchestrat is a macOS workspace that coordinates AI coding agents by providing shared tasks, task ownership, versioned Markdown documents, and live activity tracking. It integrates with existing coding tools like Claude Code and Cursor through MCP and APIs, allowing multiple agents to collaborate while users retain control and review authority. The platform is in private early access with public pricing to be announced.
OpenPrompt is a unified dashboard that coordinates multiple AI coding tools including Cursor, Claude Code, and Codex, routing tasks to appropriate models based on cost and complexity while maintaining a history of all code changes. It acts as a control layer above existing workflows without replacing them, available for individual developers, teams, or enterprise deployment.
Grafana's agento11y tool monitors multiple coding agents including Cursor, Claude Code, and Copilot CLI to track usage, costs, and tokens in a centralized dashboard. Installation is available via curl, Homebrew, Go, or prebuilt binaries, with configuration through agento11y login to store sessions locally or in Grafana Cloud.