daggerhashimoto/openclaw-nerve

What it solves

Nerve provides a comprehensive control plane and user interface for OpenClaw, moving beyond simple chat threads to a full operating surface. It solves the problem of managing multiple AI agents, their individual workspaces, memories, and configurations, which often become cluttered or invisible in a standard chat interface.

How it works

Nerve acts as a frontend "cockpit" that sits between the user's browser and the OpenClaw Gateway. It uses a combination of WebSockets (WS) for proxied gateway communication, Server-Sent Events (SSE) for real-time file watching and sync, and REST APIs for managing files, memories, and TTS. This allows users to interact with a fleet of agents through a unified interface while the OpenClaw engine handles the underlying logic.

Who it’s for

It is designed for users operating a fleet of AI agents who need high visibility, granular control, and coordination tools like task boards and session trees rather than just a conversational interface.

Highlights

• Fleet Control: Manage multiple agents, each with its own identity, skills, and memory, from a single interface.
• Integrated Voice: Built-in push-to-talk, wake word flows, local Whisper transcription, and multiple TTS providers.
• Agent Operating Context: Live inspection and editing of agent workspaces, memory, and configurations.
• Operational Tools: Includes cron scheduling, kanban workflows, session trees, and a proposal inbox.
• Rich Output: Renders charts, diffs, syntax-highlighted code, and structured tool outputs instead of plain text blocks.
• Observability: Real-time tracking of token usage, costs, and context pressure.