MeshPilot

The second major feature group is the Visual Kanban system wired to real task state. This task board displays the progress of each project task, linked directly to the actual terminal commands and file changes made by agents. The Kanban board is not just a static list; it is actively synchronized with the agent's actions. How it works: when an agent finishes a command or modifies a file, the Kanban card updates automatically, reflecting the new state. This provides a clear, visual overview of what has been done and what remains. If an agent completes a task, the card moves to "Done" automatically. The benefit is that developers can track complex workflows at a glance, without reading logs. It makes coordination with multiple agents manageable, as each task's status is transparent and reviewable.

The third feature group includes MeshMemory and the Pilot Assistant. MeshMemory is an AI-curated semantic memory layer that automatically saves context, notes, and task progress in the background. This gives the assistant a persistent, deep understanding of your codebase across sessions. The Pilot Assistant provides on-demand help and guidance. Additionally, MeshUtility, a free open-source desktop app, brings instant voice dictation and global AI prompt rewriting into a single widget for any text field on your machine. MeshUtility can be used anywhere on your desktop, not just within MeshPilot, making it a versatile tool for writing code comments, documentation, or even emails. The MeshPrompt rewriting feature uses AI to improve dictated or typed text, offering suggestions for clarity and tone. It supports local and cloud transcription with user-controlled privacy, and includes push-to-talk or toggle mode with configurable hotkeys. This integration expands MeshPilot's value beyond coding into everyday productivity.

The overall workflow in MeshPilot begins with setting up a project workspace. Users define tasks on the Visual Kanban board, then launch autonomous CLI agents that operate in persistent terminals. The agents execute commands, write code, and manage files according to the tasks, all visible in real time. Developers can watch the agents work and step in to provide input or corrections when necessary. MeshMemory automatically records the context of each session, including commands run and decisions made, building a knowledge base that agents and users can reference later. This workflow is iterative: as agents complete tasks, the Kanban board updates, allowing the developer to review the output and define new tasks. The persistent terminals mean that agents can maintain state across multiple commands, enabling complex, multi-step processes. For example, an agent could set up a database, run migrations, and deploy code in a single session without losing context. This streamlined approach reduces the time spent on manual setups and allows developers to focus on higher-level decisions.

Concrete use cases for MeshPilot include a solo developer building a web application: they can set up tasks for backend API, frontend components, and deployment scripts, then let agents handle the coding while they review progress in the Kanban board. Another scenario is a team working on a microservices project: each agent can work on a different service in separate terminals, with shared memory ensuring consistency. Non-developers can use MeshUtility for voice dictation to write code or documentation, leveraging MeshPrompt rewriting to improve text. For debugging, a developer can create a task to investigate a bug, launch an agent to run tests and log analysis, and watch the results in the live browser preview. The agent can then propose a fix, which the developer can review and approve. The outcome is faster development cycles, reduced manual terminal work, and a clear record of what was done. Users report being able to ship features more quickly by offloading routine tasks to AI.

MeshPilot targets solo developers, indie builders, and software engineering teams on Windows, macOS, and Linux. Its pricing includes a Starter plan at $5/month with full MeshConsole access, MeshMemory local, Pilot Assistant, and MeshUtility. The Plus plan at $20/month (or $10/month with code FOREVER) adds MeshMemory Cortex, hosted context sync, priority support, and early access to new features. The Starter plan is designed for individuals getting started, while the Plus plan offers advanced memory and sync features for power users. A flexible credits system is also planned for extra usage, allowing users to top up on demand, scaling with heavier workspace and swarm runs. All plans come with a 7-day free trial. The product is currently in early access, signups are open. MeshPilot's open-source component, MeshUtility, is available on GitHub for those who want to examine or contribute to its code. In summary, MeshPilot is the first dedicated agentic development environment that brings terminal, task, and agent management together, enabling a new paradigm of AI-assisted software development.