A multi-agent orchestration system for OpenClaw that models AI agent collaboration on the Tang-dynasty Three Provinces and Six Ministries structure, with a review gate and live kanban dashboard.
daggerhashimoto/openclaw-nerve
A self-hosted web cockpit for OpenClaw agents, adding voice chat, a kanban board, workspace file editing, session views and usage tracking.
- Stars
- 871
- Forks
- 148
- Open issues
- 32
- Last commit
- 16 days ago
What is it?
Nerve is a browser-based interface that sits in front of an OpenClaw gateway and gives operators a single place to run and inspect multiple agents. Each agent can have its own workspace, subagents, memory, identity, soul and skills, which can be viewed and edited live from the UI.
Beyond chat, Nerve adds an operating layer: cron runs appear as separate sessions, agent work can be delegated to a kanban board and reviewed, and token usage, cost and context pressure are tracked during long tasks. Agent replies can include real charts, diffs, previews and syntax-highlighted code.
Voice is built in, with push-to-talk, wake word flows, local Whisper transcription and several TTS providers. It runs locally by default, can be exposed on a network with password authentication, and supports access via LAN or Tailscale.
Key features
- 01Manage multiple agents from one control plane, each with its own workspace and memory
- 02Push-to-talk and wake word voice input with local Whisper transcription and multiple TTS providers
- 03Per-agent file browser and tabbed editor for workspace, memory and config files
- 04Session tree, cron scheduling and kanban task board for agent work
- 05Token usage, cost tracking and context meter displayed during runs
- 06Charts, diffs and syntax-highlighted code rendered in agent responses
Use cases
- Operating several OpenClaw agents and switching between their workspaces and memory
- Delegating tasks to agents on a kanban board and reviewing their output
- Monitoring cron-driven agent runs and subagent activity without leaving the main thread
- Talking to a self-hosted agent by voice from a browser or phone
Topics
FAQ
- What is openclaw-nerve?
- A self-hosted web cockpit for OpenClaw agents, adding voice chat, a kanban board, workspace file editing, session views and usage tracking. Nerve is a browser-based interface that sits in front of an OpenClaw gateway and gives operators a single place to run and inspect multiple agents. Each agent can have its own workspace, subagents, memory, identity, soul and skills, which can be viewed and edited live from the UI.
- Is openclaw-nerve free to use? What license does it have?
- openclaw-nerve is open source under the MIT license. Check the license terms before using it in commercial products.
- Is openclaw-nerve still maintained?
- Actively maintained. The last commit to openclaw-nerve was on Oct 6, 2026.
- What can I use openclaw-nerve for?
- Operating several OpenClaw agents and switching between their workspaces and memory Delegating tasks to agents on a kanban board and reviewing their output Monitoring cron-driven agent runs and subagent activity without leaving the main thread Talking to a self-hosted agent by voice from a browser or phone
- Where can I get openclaw-nerve?
- The source code and installation instructions are on GitHub: https://github.com/daggerhashimoto/openclaw-nerve
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An MCP voice layer that lets you have spoken conversations with Claude Code and other MCP-capable agents, using cloud or local speech services.
A skill package that gives AI agents real-time web search, vertical domain search, batch search, and full-page content extraction through a CLI.
A thin local proxy that lets Hermes Agent, OpenClaw and OpenCode share one WeChat iLink account, switchable with chat commands.
MCP server for AI coding agents that keeps raw tool output out of the context window and persists session state across compaction.
A Claude Code, OpenCode and Codex skill that runs goal-driven modify-verify-keep/discard loops, inspired by Karpathy's autoresearch.