English · 简体中文
Your long-term personal AI partner.
NextClaw's mission is to be a good long-term partner to you. We are building toward a 1.0 that understands you, takes useful initiative, and continuously improves itself. Today, you can use it to discuss questions, analyze material, complete work, and follow up through scheduled tasks.
From one request to a usable result, without losing the work in between.
Download & Install · Explore Use Cases · Read the Docs
Open source · Local-first · macOS, Windows, Linux, Docker, and cloud VMs
The goal for 1.0 is to deliver a long-term AI partner that understands you, takes useful initiative, and continuously improves itself. Initiative includes independent analysis, planning, sharing relevant information, and starting a conversation, as well as acting within your authorization.
| Stage | Planned delivery |
|---|---|
| Understand you and your goals | Organize user background, review and correct memory, manage personal goals and plans, and use current understanding in new conversations and background work |
| Proactive analysis and conversation | Identify ongoing interests, analyze and plan in the background, control timing, start and continue discussions, and act and follow up when appropriate |
| Continuous self-improvement | Reflect on results and feedback, apply experience, inspect its effects, and correct, disable, or reverse unsuitable changes |
| Reliable long-term use | Improve setup, capability status, recovery, and upgrades; pass core acceptance cases and continuous user trials |
This is planned scope, building on existing memory, scheduling, conversations, and skills. Full acceptance testing is not complete, and no release date is set. The sections below describe how to use the current product.
Repository planning documents (Chinese): Feature development table · 1.0 definition and acceptance cases · Current capabilities and gaps
|
Start with the goal Ask for a report, analysis, file operation, small app, or recurring task. NextClaw organizes the tools and steps behind it. |
Keep context and results together Conversations, local files, web research, generated documents, and follow-up work stay in the same task. |
Choose how the work runs Use Native, Codex, Claude Code, OpenCode, or Hermes, then run locally, on a NAS, or on a server you control. |
NextClaw also treats the nextclaw CLI as a first-class way to work. Many core operations can be used from a terminal, script, CI job, or another Agent, with machine-readable output where supported.
New installations are ready for the first task without entering an API key. The built-in free trial uses a public gateway; limits and models may change, and sensitive or confidential data should not be sent through it.
- Research and compare — collect pages, notes, and references, then turn them into a brief, source list, or comparison table.
- Analyze and visualize data — gather data from websites, CSV files, or spreadsheets, clean it, draw charts, and write the conclusion.
- Draft useful documents — shape source material and rough notes into reports, articles, proposals, release notes, or weekly updates.
- Process local files — inspect, rename, extract, classify, and summarize documents without losing the task context.
- Build small tools for yourself — turn a repeated job into a script, local app, dashboard, or reusable workflow.
- Keep recurring work moving — receive requests from chat apps, run scheduled briefs or checks, and send results back to the right channel.
Start with source material, let the Agent organize and visualize it, then inspect the result beside the conversation. Source files and project documents remain available in the same workspace.
When scheduled work, a background Agent, or a long-running monitor finishes, NextClaw can deliver the report to your inbox. Read it later, manage it with the rest of your results, or continue the conversation with the full context.
Keep an Agent's identity, workspace, memory, and skills, then run the task with Native, Codex, Claude Code, OpenCode, or Hermes. Choose the runtime when starting a task and keep the rest of the workspace unchanged.
Open code, Markdown, HTML, Word, Excel, and PowerPoint without losing the task that produced them.
Build a page with an Agent, run it beside the conversation, and keep it as a Panel App you can open and improve later.
|
Dedicated Agents Give different kinds of work their own role, memory, skills, runtime, and workspace. 
|
Image generation Generate an image, keep the local file, and continue using it in the same task. 
|
|
Messaging channels Connect Weixin, Feishu/Lark, QQ, and other channels to the Agents running on your machine. 
|
Scheduled work Run recurring briefs, checks, and other tasks on a schedule you control. 
|
|
Skills and references Install skills while keeping their documentation open in the right-side Doc Browser. 
|
Model providers Use built-in providers or add your own OpenAI-compatible endpoint and models. 
|
The desktop app is the easiest way to start on macOS, Windows, or Linux.
Download the latest stable desktop release
Install Node.js LTS first, then run:
npm install -g nextclaw
nextclaw startOpen http://127.0.0.1:55667 and start with the built-in free-trial model. Connect your own provider when you need it.
If npm is unavailable, install or reinstall Node.js LTS and reopen the terminal. On a remote host, port 55667 serves plain HTTP. Use it directly only for a quick check; terminate HTTPS with Nginx or Caddy for regular access.
nextclaw stopFor a long-running server or cloud VM deployment:
curl -fsSL https://nextclaw.io/install-docker.sh | bashSee the Docker deployment guide for reverse proxy, domain, and remote access setup. You can compare every supported path on the download and install page.
For the current tested server baseline, idle measurements, and factors that increase memory during active work, see Runtime Resource Usage.
- Models — built-in free-trial access, plus OpenRouter, OpenAI, Anthropic, Gemini, DeepSeek, MiniMax, Moonshot, DashScope, Zhipu, AiHubMix, vLLM, and custom OpenAI-compatible endpoints.
- Messaging channels — Weixin, Feishu/Lark, QQ, DingTalk, WeCom, Telegram, Discord, Slack, WhatsApp, and email.
- Capabilities — skills, MCP servers, CLI tools, browser control, local files, Panel Apps, and scheduled tasks.
- Local control — configuration, conversations, and credentials stay in the environment you control. Connected providers and channels receive the data you send through them.
From the repository root:
pnpm install
pnpm dev startThe development stack prints its local URLs in the terminal and uses ~/.nextclaw by default. Set NEXTCLAW_HOME=/path/to/home to use an isolated data directory.
To run only one side:
pnpm dev:backend
pnpm dev:frontendSee the developer command reference for local source-runtime checks, the manual runtime-update harness, platform stacks, Docker, and validation commands.
To refresh the repository and website screenshot set:
pnpm run screenshots:refresh- Getting Started
- Configuration
- Model Selection
- Commands
- Feishu Setup
- Vision
- 1.0 Roadmap
- Product Updates
Repository planning: Roadmap · TODO
- GitHub Issues
- WeChat group: scan the QR code below.
Suggestions are welcome through GitHub Issues: report a bug, propose an idea or improvement, or browse work open to outside contributors. Please read CONTRIBUTING.md before implementing anything: an Issue needs help wanted plus explicit maintainer confirmation or assignment before a pull request is opened.
NextClaw was inspired by these projects:
- OpenClaw — inspired NextClaw's early exploration of a full-stack AI assistant.
- NanoBot — demonstrated how a small agent framework can remain useful and extensible.
