create-openclaw-bot 5.0.0 → 5.0.3

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,237 +1,321 @@
1
- <div align="center">
2
-
3
- # 🦞 OpenClaw Setup
4
-
5
- <p align="center">
6
- <a href="https://github.com/tuanminhhole/openclaw-setup/releases"><img src="https://img.shields.io/badge/RELEASE-v5.0.0-0EA5E9?style=for-the-badge" alt="Version 5.0.0" /></a>
7
- <a href="https://github.com/tuanminhhole/openclaw-setup?tab=MIT-1-ov-file"><img src="https://img.shields.io/badge/LICENSE-MIT-success?style=for-the-badge" alt="MIT License" /></a>
8
- <a href="https://www.npmjs.com/package/create-openclaw-bot"><img src="https://img.shields.io/npm/v/create-openclaw-bot?style=for-the-badge&label=CLI&color=2563EB&logo=npm&logoColor=white" alt="NPM Version" /></a>
9
- <a href="https://github.com/tuanminhhole/openclaw-setup/stargazers"><img src="https://img.shields.io/github/stars/tuanminhhole/openclaw-setup?style=for-the-badge&color=eab308&logo=github&logoColor=white" alt="GitHub Stars" /></a>
10
- </p>
11
-
12
- <p style="margin-top: 16px;">
13
- <img src="https://flagcdn.com/24x18/gb.png" alt="English" width="24" height="18" style="vertical-align: sub;"> <strong>English</strong> ·
14
- <img src="https://flagcdn.com/24x18/vn.png" alt="Tiếng Việt" width="24" height="18" style="vertical-align: sub;"> <a href="README.vi.md">Tiếng Việt</a>
15
- </p>
16
-
17
- An interactive <strong>CLI tool</strong> and <strong>Setup Wizard</strong> to deploy your own free AI Bot on Telegram or Zalo in minutes.
18
-
19
- <a href="https://github.com/tuanminhhole/openclaw-setup">
20
- <img src="https://raw.githubusercontent.com/tuanminhhole/openclaw-setup/main/docs/preview.png" alt="OpenClaw Setup Hero Image" width="100%" style="border-radius: 8px; margin: 16px 0; border: 1px solid #333;" />
21
- </a>
22
-
23
- </div>
24
-
25
- ---
26
-
27
- ## 🆕 What's new in v5.0.0 — Gemma 4 Support
28
-
29
- > **Google just dropped Gemma 4 on April 2, 2026. OpenClaw ships support immediately.**
30
-
31
- - 🧠 **Gemma 4 (Google, Open-weights)** — The newest open model from Google DeepMind. Free, powerful, Apache 2.0. Available in 3 sizes: `gemma4:4b` (6 GB), `gemma4` (10 GB), `gemma4:27b` (18 GB).
32
- - 🐳 **Auto Ollama Sidecar** — No need to install Ollama separately. When you pick Local Ollama, the setup generates a full `docker-compose.yml` with an **Ollama container** that auto-downloads your chosen model on first launch.
33
- - 🆓 **Zero cost, zero API key** Gemma 4 runs 100% locally in Docker. Perfect for privacy-first setups.
34
-
35
- <details>
36
- <summary><b>Previous: What's new in OpenClaw v4</b></summary>
37
-
38
- - 🚀 **Zero-Config `npx` Deployment**: Interactive CLI `npx create-openclaw-bot` and Web Wizard builds the entire Docker workspace dynamically.
39
- - 🔀 **Simplified 9Router Smart Routing**: Auto load-balances across Anthropic, Codex, Gemini, Qwen. No API keys needed.
40
- - 🌐 **Instant Browser Automation**: Built-in `.bat` and `.sh` scripts for Chrome Debug Mode on Windows/macOS/Linux.
41
- - 🧹 **Zero-Clutter Repository**: Generates exactly what you need, nothing more.
42
-
43
- </details>
44
-
45
- ---
46
-
47
- ## ✨ Features
48
-
49
- - 🤖 **Multi-channel** — Telegram, Zalo Bot API, or Zalo Personal
50
- - 🧠 **Multi-provider AI** — Google Gemini, Claude, GPT-4o/Codex, OpenRouter, Ollama, 9Router
51
- - 🧩 **Skills** — Web Search, Browser Automation, Memory, RAG, Code Interpreter, Image Gen...
52
- - 🔌 **Plugins** — Voice Call, Matrix, MS Teams, Nostr
53
- - 🔀 **9Router Proxy** — Auto-route your bot's traffic effectively!
54
- - 🧙 **Setup Wizard** — Beautiful interactive shell (CLI) or dark-themed Web UI
55
- - 🆓 **100% Free** — Google Gemini API free tier + Docker
56
- - 🔒 **Secure** — API keys stay on your machine, never exposed
57
- - **3 minutes** — From zero to working AI bot
58
-
59
- ---
60
-
61
- ## 📋 Prerequisites
62
-
63
- You need **3 things** (all free):
64
-
65
- 1. **Docker Desktop** — [Download here](https://www.docker.com/products/docker-desktop/)
66
- 2. **AI API Key** See [Supported Providers](#-supported-providers) (Gemini is free!)
67
- 3. **Bot Token** — See [Supported Channels](#-supported-channels)
68
-
69
- ---
70
-
71
- ## 🧠 Supported Providers
72
-
73
- - **Google Gemini** (Gemini 2.5 Flash/Pro, 3.0 Flash) — 🆓 Free — [Get Key](https://aistudio.google.com/apikey)
74
- - **Anthropic Claude** (Sonnet 4, Opus 4, Haiku 3.5) — 💰 Paid — [Get Key](https://console.anthropic.com/settings/keys)
75
- - **OpenAI / Codex** (GPT-4o, o3, Codex Mini) 💰 Paid [Get Key](https://platform.openai.com/api-keys)
76
- - **OpenRouter** (Many free & paid models) — 🆓/💰 — [Get Key](https://openrouter.ai/keys)
77
- - **Ollama / Gemma 4** (local models, no API key) 🏠 Local — **Auto-runs in Docker, no manual install**
78
- - **9Router** (Auto-routes to best provider) — 🔀 Proxy — [Docs](https://github.com/decolua/9router)
79
-
80
- > 🔀 **9Router** runs alongside OpenClaw in Docker. After `docker compose up`, open `localhost:20128/dashboard` to login via OAuth. No API keys needed!
81
-
82
- ---
83
-
84
- ## 🔌 Supported Channels
85
-
86
- - **Telegram** (✅ Official) Search **@BotFather** on Telegram `/newbot` → Copy token.
87
- - **Zalo Bot API** (✅ Official) Go to [developers.zalo.me](https://developers.zalo.me) Create bot → Copy token.
88
- - **Zalo Personal** (⚠️ Unofficial) — Scan QR code after Docker setup (no token needed).
89
-
90
- > ⚠️ **Zalo Personal** uses an unofficial API. Your account may be restricted. Use a secondary account.
91
-
92
- ---
93
-
94
- ## 🚀 Quick Start
95
-
96
- ### 1️⃣ Option A Interactive CLI (`npx`) [Recommended]
97
-
98
- The fastest way to install OpenClaw is using the interactive NPM package.
99
-
100
- 1. **Open your Terminal (or Command Prompt).**
101
- 2. **Run the command:**
102
- ```bash
103
- npx create-openclaw-bot
104
- ```
105
- 3. **Follow the interactive prompts** to configure your bot, keys, and paths!
106
-
107
- <details>
108
- <summary><b>2️⃣ Option B — GUI Setup Wizard (Web UI)</b></summary>
109
- <br>
110
-
111
- 1. **Clone this repo:**
112
-
113
- ```bash
114
- git clone https://github.com/tuanminhhole/openclaw-setup.git
115
- cd openclaw-setup
116
- ```
117
-
118
- 2. **Open `index.html`** in your browser — the Setup Wizard will guide you through:
119
- - Choose your channel (Telegram / Zalo)
120
- - Configure bot name, personality, AI model
121
- - Generate all config files and `setup-openclaw.bat`
122
-
123
- </details>
124
-
125
- ### 3️⃣ Option CAI Agent (Antigravity)
126
-
127
- 1. Open [Antigravity IDE](https://antigravity.dev/)
128
- 2. Open this repo as workspace
129
- 3. Paste into chat:
130
- ```text
131
- Read SETUP.md and set up OpenClaw v5.0.0 for me.
132
- My bot token is X, my 9Router proxy doesn't need a key.
133
- My project folder: <YOUR_PATH>
134
- ```
135
-
136
- ---
137
-
138
- ## 📁 Repo Structure
139
-
140
- ```
141
- index.html ← Setup Wizard UI
142
- style.css ← Wizard styles
143
- setup.js ← Wizard logic
144
- CHANGELOG.md ← Version history
145
- README.md ← You're here (English)
146
- README.vi.md ← Hướng dẫn tiếng Việt
147
- SETUP.md ← Technical guide for AI (English)
148
- SETUP.vi.md ← Hướng dẫn kỹ thuật cho AI (Tiếng Việt)
149
- ```
150
-
151
- ---
152
-
153
- ## FAQ
154
-
155
- <details>
156
- <summary><b>Is it really free?</b></summary>
157
- Yes. Docker, Google Gemini API (free tier), and Telegram/Zalo bots are all free.
158
- </details>
159
-
160
- <details>
161
- <summary><b>Where does the bot run?</b></summary>
162
- On your computer inside a Docker container. When your PC is off, the bot is off.
163
- </details>
164
-
165
- <details>
166
- <summary><b>How do I stop/restart?</b></summary>
167
-
168
- ```bash
169
- docker compose down # Stop
170
- docker compose up -d # Start
171
- ```
172
-
173
- </details>
174
-
175
- <details>
176
- <summary><b>Can I customize the bot?</b></summary>
177
- Yes! Use the Setup Wizard to configure name, personality, and language. Or edit the YAML files directly.
178
- </details>
179
-
180
- <details>
181
- <summary><b>Is it safe?</b></summary>
182
- Your API keys stay on your machine only. SETUP.md includes strict security rules that the AI must follow.
183
- </details>
184
-
185
- <details>
186
- <summary><b>Can I switch channels later?</b></summary>
187
- Yes! Re-run the Setup Wizard or manually edit <code>.openclaw/openclaw.json</code> to change the channel config.
188
- </details>
189
-
190
- <details>
191
- <summary><b>What is 9Router?</b></summary>
192
- 9Router is a community open-source AI proxy that runs in Docker alongside your bot. Instead of API keys, you login via OAuth on the 9Router dashboard (<code>localhost:20128</code>). It auto-routes requests to the best available AI provider.
193
- </details>
194
-
195
- <details>
196
- <summary><b>What's the difference between Skills and Plugins?</b></summary>
197
- <strong>Skills</strong> add agent capabilities (Web Search, Browser, Memory, etc.) — installed via <code>openclaw skills install</code> from ClawHub.<br>
198
- <strong>Plugins</strong> add channels or runtime extensions (Voice Call, Matrix, etc.) installed via <code>openclaw plugins install</code> from npm.
199
- </details>
200
-
201
- ---
202
-
203
- ## 🔗 Links
204
-
205
- - [OpenClaw Docs](https://openclaw.ai/docs)
206
- - [Google AI Studio](https://aistudio.google.com/)
207
- - [Telegram BotFather](https://t.me/BotFather)
208
- - [Zalo Bot Platform](https://developers.zalo.me)
209
- - [Docker Desktop](https://www.docker.com/products/docker-desktop/)
210
-
211
- ---
212
-
213
- ## 📈 Star History
214
-
215
- <div align="center">
216
-
217
- [![Star History Chart](https://starchart.cc/tuanminhhole/openclaw-setup.svg?variant=adaptive)](https://starchart.cc/tuanminhhole/openclaw-setup)
218
-
219
- </div>
220
-
221
- ---
222
-
223
- ## 🙏 Acknowledgments
224
-
225
- - [OpenClaw](https://openclaw.ai) — AI Gateway framework
226
- - [9Router](https://github.com/decolua/9router) — Open-source AI proxy (OAuth-based, no API keys)
227
- - [Playwright](https://playwright.dev) — Browser automation engine
228
- - [ClawHub](https://clawhub.com) — Skills registry
229
- - [TheSVG](https://thesvg.org) — High-quality SVG brand icons
230
-
231
- ---
232
-
233
- <div align="center">
234
-
235
- Made with 🦞 by [Kent](https://github.com/tuanminhhole)
236
-
237
- </div>
1
+ <div align="center">
2
+
3
+ # 🦞 OpenClaw Setup
4
+
5
+ <p align="center">
6
+ <a href="https://github.com/tuanminhhole/openclaw-setup/releases"><img src="https://img.shields.io/badge/RELEASE-v5.0.1-0EA5E9?style=for-the-badge" alt="Version 5.0.1" /></a>
7
+ <a href="https://github.com/tuanminhhole/openclaw-setup?tab=MIT-1-ov-file"><img src="https://img.shields.io/badge/LICENSE-MIT-success?style=for-the-badge" alt="MIT License" /></a>
8
+ <a href="https://www.npmjs.com/package/create-openclaw-bot"><img src="https://img.shields.io/npm/v/create-openclaw-bot?style=for-the-badge&label=CLI&color=2563EB&logo=npm&logoColor=white" alt="NPM Version" /></a>
9
+ <a href="https://github.com/tuanminhhole/openclaw-setup/stargazers"><img src="https://img.shields.io/github/stars/tuanminhhole/openclaw-setup?style=for-the-badge&color=eab308&logo=github&logoColor=white" alt="GitHub Stars" /></a>
10
+ </p>
11
+
12
+ <p style="margin-top: 16px;">
13
+ <img src="https://flagcdn.com/24x18/gb.png" alt="English" width="24" height="18" style="vertical-align: sub;"> <strong>English</strong> ·
14
+ <img src="https://flagcdn.com/24x18/vn.png" alt="Tiếng Việt" width="24" height="18" style="vertical-align: sub;"> <a href="README.vi.md">Tiếng Việt</a>
15
+ </p>
16
+
17
+ An interactive **CLI tool** and **Setup Wizard** to deploy your own free AI Bot on Telegram or Zalo in minutes — supports **Windows, macOS, Ubuntu, and VPS**.
18
+
19
+ <a href="https://github.com/tuanminhhole/openclaw-setup">
20
+ <img src="https://raw.githubusercontent.com/tuanminhhole/openclaw-setup/main/docs/preview.png" alt="OpenClaw Setup Hero Image" width="100%" style="border-radius: 8px; margin: 16px 0; border: 1px solid #333;" />
21
+ </a>
22
+
23
+ </div>
24
+
25
+ ---
26
+
27
+ ## 🆕 What's new in v5.0.1
28
+
29
+ - 💻 **OS-First Setup** Step 1 is now choosing your OS (Windows, macOS, Ubuntu, VPS). All scripts, configs, and instructions are generated to match.
30
+ - 🧠 **Gemma 4 — 4 sizes** — `gemma4:e2b` (~4 GB), `gemma4:e4b` (~8 GB), `gemma4:26b` (~18 GB), `gemma4:31b` (~24 GB). Auto-pulled on first launch.
31
+ - 📄 **Consistent script naming** — Scripts are generated per OS: `setup-openclaw-docker-win.bat`, `setup-openclaw-docker-macos.sh`, `setup-openclaw-vps.sh`... No more environment confusion.
32
+ - 🌐 **Native Install for Ubuntu/VPS** — No Docker needed. Both CLI and Web Wizard support running directly with PM2 less RAM overhead, stable 24/7.
33
+ - 🤖 **Multi-Bot Deployment** Deploy up to **5 independent Telegram bots** in one setup. Each bot runs in its own isolated workspace (`bot1/`, `bot2/`...) with its own token, name, slash command, and AI personality.
34
+ - 💬 **Department Room Model** — Add all bots to one Telegram group and they behave like a professional team. Bots stay silent by default, only responding when @mentioned or their slash command is used — no spam, no chaos.
35
+ - 🔗 **Group ID Helper** — "Get Group ID" button in Web Wizard + step-by-step guide in CLI. Opens @userinfobot directly — no documentation hunting needed.
36
+
37
+ <details>
38
+ <summary><b>Previous: What's new in v5.0.0</b></summary>
39
+
40
+ - 🤖 **Gemma 4 (Google, Open-weights)** The newest model from Google DeepMind. Free, powerful, Apache 2.0.
41
+ - 🐳 **Auto Ollama Sidecar** — No manual Ollama install needed. Setup generates a `docker-compose.yml` with an Ollama container that auto-downloads your model on first launch.
42
+ - 🆓 **Zero cost, zero API key** — Gemma 4 runs 100% locally in Docker.
43
+
44
+ </details>
45
+
46
+ ---
47
+
48
+ ## ✨ Features
49
+
50
+ - 🤖 **Multi-channel** — Telegram (single or multi-bot), Zalo Bot API, or Zalo Personal
51
+ - 🧑‍🤝‍🧑 **Multi-bot team** — Run up to 5 Telegram bots simultaneously. Bots share a workspace, collaborate on tasks, and coordinate in a group chat using the Department Room Model
52
+ - 🧠 **Multi-provider AI** — Google Gemini, Claude, GPT-4o, OpenRouter, Ollama (local), 9Router
53
+ - 🧩 **Skills** — Web Search, Browser Automation, Memory, RAG, Code Interpreter, Image Gen
54
+ - 🔌 **Plugins** — Voice Call, Matrix, MS Teams, Nostr
55
+ - 🔀 **9Router** — Free AI proxy. OAuth login, no API keys. Supports Claude Code, Codex, Gemini CLI.
56
+ - 🧙 **Setup Wizard** — 5-step visual web UI (`index.html`). No terminal required.
57
+ - 💻 **Interactive CLI** — `npx create-openclaw-bot` best for Ubuntu, VPS, engineers.
58
+ - 🆓 **100% Free to start** — 9Router + Gemini free tier requires zero spending
59
+ - 🔒 **Private** — API keys stay on your machine, never sent anywhere
60
+ - ⚡ **Fast** — From zero to working bot in under 5 minutes
61
+
62
+ ---
63
+
64
+ ## 🗺️ Choose your path
65
+
66
+ > **Not sure which method to use?** The table below has you covered:
67
+
68
+ | Who you are | Environment | Recommended path |
69
+ | -------------------------- | --------------- | ------------------------------------- |
70
+ | Not familiar with terminal | Windows / macOS | **Web Wizard** (`index.html`) |
71
+ | Not familiar with terminal | Ubuntu Desktop | **Web Wizard** → choose Native |
72
+ | Comfortable with terminal | Ubuntu / VPS | **CLI** (`npx create-openclaw-bot`) |
73
+ | Want full automation | Anywhere | **AI Agent** (Antigravity + SETUP.md) |
74
+
75
+ ### 1️⃣ Option AWeb Wizard (No terminal required)
76
+
77
+ Best for **Windows and macOS**. No command line at all.
78
+
79
+ 1. [Download ZIP](https://github.com/tuanminhhole/openclaw-setup/archive/refs/heads/main.zip) or clone this repo.
80
+ 2. Open `index.html` in your browser.
81
+ 3. Follow the **5-step wizard**:
82
+ - **Step 1:** Choose your OS (Windows / macOS / Ubuntu / VPS)
83
+ - **Step 2:** Choose your bot channel (Telegram / Zalo)
84
+ - **Step 3:** Choose your AI provider and model
85
+ - **Step 4:** Enter bot token and configure settings
86
+ - **Step 5:** Download your script and run it done!
87
+ 4. The downloaded script automatically installs everything needed (9Router, Ollama, Docker, etc.) based on your choices.
88
+
89
+ > **Docker or no Docker?**
90
+ >
91
+ > - **Windows / macOS** → Use **Docker** (fully isolated, easy to manage)
92
+ > - **Ubuntu / VPS** → Use **Native (no Docker)** (less RAM, more stable)
93
+
94
+ ### 2️⃣ Option B — Interactive CLI (`npx`)
95
+
96
+ Best for **engineers, Ubuntu Desktop, VPS**. Fastest and most powerful.
97
+
98
+ ```bash
99
+ npx create-openclaw-bot
100
+ ```
101
+
102
+ Run in your terminal → follow the interactive prompts → startup script is generated automatically.
103
+
104
+ > Requires: **Node.js 20 LTS** or newer. Check: `node -v`
105
+
106
+ <details>
107
+ <summary><b>3️⃣ Option C — AI Agent (Antigravity)</b></summary>
108
+ <br>
109
+
110
+ 1. Open [Antigravity IDE](https://antigravity.dev/)
111
+ 2. Open this repo as your workspace
112
+ 3. Paste into chat:
113
+ ```
114
+ Read SETUP.md and set up OpenClaw v5.0.1 for me.
115
+ My bot token is X. Use 9Router (no API key).
116
+ My project folder: <YOUR_PATH>
117
+ ```
118
+
119
+ </details>
120
+
121
+ ---
122
+
123
+ ## 📋 Prerequisites
124
+
125
+ ### Without Docker (Nativerecommended for Ubuntu/VPS)
126
+
127
+ | Requirement | Notes |
128
+ | ----------------------- | ------------------------------------------------------------------------- |
129
+ | **Node.js 20 LTS** | [Download](https://nodejs.org/) · Check: `node -v` |
130
+ | **An AI provider** | 9Router (free) or Gemini/Claude/GPT-4o |
131
+ | **Bot Token** | From Telegram BotFather or Zalo Developer |
132
+ | **Ollama** _(optional)_ | Only if you want to run Gemma 4 locally · [Download](https://ollama.com/) |
133
+
134
+ ### With Docker (recommended for Windows/macOS)
135
+
136
+ | Requirement | Notes |
137
+ | ------------------------------- | --------------------------------------------------------------------------------------------- |
138
+ | **Node.js 20 LTS** | [Download](https://nodejs.org/) · Check: `node -v` |
139
+ | **Docker Desktop + Compose V2** | [Download](https://www.docker.com/products/docker-desktop/) · Check: `docker compose version` |
140
+ | **An AI provider** | 9Router runs as a sidecar container — no separate install needed |
141
+ | **Bot Token** | From Telegram BotFather or Zalo Developer |
142
+
143
+ ---
144
+
145
+ ## 🧠 Supported Providers
146
+
147
+ | Provider | Cost | API Key | Notes |
148
+ | -------------------- | ------------ | -------- | ----------------------------------------------------------------------------------------- |
149
+ | **9Router** | 🆓 Free | ❌ OAuth | Recommended for beginners. Auto-routes to best model. Supports Claude CLI, Codex, Gemini. |
150
+ | **Google Gemini** | 🆓 Free tier | ✅ Yes | High quality. Very generous free tier. |
151
+ | **Ollama / Gemma 4** | 🏠 Free | ❌ No | Runs 100% offline. Auto-pulls model on first start. |
152
+ | **Anthropic Claude** | 💰 Paid | ✅ Yes | Best reasoning and writing quality. |
153
+ | **OpenAI / Codex** | 💰 Paid | ✅ Yes | GPT-4o, Codex Mini. |
154
+ | **OpenRouter** | 🆓/💰 Mixed | ✅ Yes | Many models under one key. Some are free. |
155
+
156
+ > 🔀 **9Router v0.3.75+** adds lossless passthrough for Claude Code, Codex, Gemini CLI, and Antigravity — meaning these AI tools can use 9Router as their endpoint without any data loss. See [docs/ai-providers.md](docs/ai-providers.md) for setup details.
157
+
158
+ ---
159
+
160
+ ## 🔌 Supported Channels
161
+
162
+ - **Telegram** (✅ Official) Search **@BotFather** `/newbot` Copy token.
163
+ - **Zalo Bot API** (✅ Official) — Go to [developers.zalo.me](https://developers.zalo.me) → Create bot → Copy token.
164
+ - **Zalo Personal** (⚠️ Unofficial) — Scan QR after setup (no token needed). Use a secondary account.
165
+
166
+ > ⚠️ **Zalo Personal** uses an unofficial API. Your account may be restricted. Use a secondary account only.
167
+
168
+ ---
169
+
170
+ ## 📁 Repo Structure
171
+
172
+ ```
173
+ index.html ← Setup Wizard UI (open in browser)
174
+ style.css ← Wizard styles
175
+ setup.js ← Wizard logic
176
+ cli.js ← Interactive CLI (npx create-openclaw-bot)
177
+ CHANGELOG.md/.vi.md ← Version history
178
+ README.md ← You're here (English)
179
+ README.vi.md ← Hướng dẫn tiếng Việt
180
+ SETUP.md/.vi.md ← Technical guide for AI Agent
181
+ docs/
182
+ install-docker.md/.vi.md Docker setup per OS
183
+ install-native.md/.vi.md ← Native/PM2 install per OS
184
+ ai-providers.md/.vi.md ← AI provider configuration
185
+ hardware-guide.md/.vi.md ← RAM planning for Ollama/Gemma 4
186
+ faq.md/.vi.md ← Frequently asked questions
187
+ ```
188
+
189
+ > **Note:** Startup scripts (`.bat`, `.sh`) are **not included** in the repo — they are generated by the Web Wizard or CLI based on your specific configuration.
190
+
191
+ ---
192
+
193
+ ## ❓ FAQ
194
+
195
+ <details>
196
+ <summary><b>Is it really free?</b></summary>
197
+
198
+ Yes. Docker, Google Gemini API (free tier), and Telegram/Zalo bot tokens are all free. You only pay if you choose a paid AI provider like Claude or GPT-4o.
199
+
200
+ </details>
201
+
202
+ <details>
203
+ <summary><b>Where does the bot run?</b></summary>
204
+
205
+ On your computer or server. With Docker it runs in a container; with Native mode it runs as a PM2-managed process. If the machine is off, the bot is off. Use a VPS for 24/7 uptime.
206
+
207
+ </details>
208
+
209
+ <details>
210
+ <summary><b>Do I need Docker?</b></summary>
211
+
212
+ No. Docker is optional. Windows/macOS users should use Docker for clean isolation. Ubuntu/VPS users should install natively with PM2 — less overhead and OpenClaw is already secure by design.
213
+
214
+ </details>
215
+
216
+ <details>
217
+ <summary><b>How do I stop/restart the bot?</b></summary>
218
+
219
+ **Docker:**
220
+
221
+ ```bash
222
+ docker compose down # Stop
223
+ docker compose up -d # Start
224
+ docker compose restart # Restart
225
+ ```
226
+
227
+ **PM2 (native):**
228
+
229
+ ```bash
230
+ pm2 stop openclaw-bot
231
+ pm2 start openclaw-bot
232
+ pm2 restart openclaw-bot
233
+ ```
234
+
235
+ </details>
236
+
237
+ <details>
238
+ <summary><b>Can I switch AI models later?</b></summary>
239
+
240
+ Yes. Re-run `npx create-openclaw-bot` in your bot folder, or edit `.openclaw/openclaw.json` directly and restart the bot.
241
+
242
+ </details>
243
+
244
+ <details>
245
+ <summary><b>Is it safe?</b></summary>
246
+
247
+ Your API keys are stored only on your machine in a local `.env` file. OpenClaw never transmits them anywhere. When using Ollama, all AI inference runs completely offline.
248
+
249
+ </details>
250
+
251
+ <details>
252
+ <summary><b>What is 9Router?</b></summary>
253
+
254
+ 9Router is an open-source AI proxy. Instead of managing API keys from multiple providers, you log in once via OAuth at `localhost:20128/dashboard`. It auto-routes requests to the best available AI model. Starting v0.3.75, it also supports lossless passthrough for Claude Code, Codex, Gemini CLI, and Antigravity.
255
+
256
+ </details>
257
+
258
+ <details>
259
+ <summary><b>What's the difference between Skills and Plugins?</b></summary>
260
+
261
+ **Skills** add agent capabilities (Web Search, Browser Automation, Memory, RAG, Code Interpreter...) — install via `openclaw skills install` from ClawHub.
262
+
263
+ **Plugins** add channels or runtime extensions (Voice Call, Matrix, MS Teams...) — install via `openclaw plugins install` from npm.
264
+
265
+ </details>
266
+
267
+ <details>
268
+ <summary><b>How much RAM do I need for Gemma 4?</b></summary>
269
+
270
+ | Model | Min RAM (Native) | Min RAM (Docker) |
271
+ | ------------ | ---------------- | ---------------- |
272
+ | `gemma4:e2b` | ~4 GB | ~5 GB |
273
+ | `gemma4:e4b` | ~8 GB | ~9 GB |
274
+ | `gemma4:26b` | ~18 GB | ~20 GB |
275
+ | `gemma4:31b` | ~24 GB | ~26 GB |
276
+
277
+ See [docs/hardware-guide.md](docs/hardware-guide.md) for full details including VPS swap setup.
278
+
279
+ </details>
280
+
281
+ ---
282
+
283
+ ## 🔗 Links
284
+
285
+ - [OpenClaw Docs](https://openclaw.ai/docs)
286
+ - [9Router](https://github.com/decolua/9router)
287
+ - [Google AI Studio](https://aistudio.google.com/)
288
+ - [Telegram BotFather](https://t.me/BotFather)
289
+ - [Zalo Developer Platform](https://developers.zalo.me)
290
+ - [Docker Desktop](https://www.docker.com/products/docker-desktop/)
291
+ - [Ollama](https://ollama.com)
292
+ - [OpenRouter](https://openrouter.ai)
293
+ - [ClawHub (Skills)](https://clawhub.com)
294
+
295
+ ---
296
+
297
+ ## 📈 Star History
298
+
299
+ <div align="center">
300
+
301
+ [![Star History Chart](https://starchart.cc/tuanminhhole/openclaw-setup.svg?variant=adaptive)](https://starchart.cc/tuanminhhole/openclaw-setup)
302
+
303
+ </div>
304
+
305
+ ---
306
+
307
+ ## 🙏 Acknowledgments
308
+
309
+ - [OpenClaw](https://openclaw.ai) — AI Gateway framework
310
+ - [9Router](https://github.com/decolua/9router) — Open-source AI proxy (OAuth-based, no API keys)
311
+ - [Playwright](https://playwright.dev) — Browser automation engine
312
+ - [ClawHub](https://clawhub.com) — Skills registry
313
+ - [TheSVG](https://thesvg.org) — High-quality SVG brand icons
314
+
315
+ ---
316
+
317
+ <div align="center">
318
+
319
+ Made with 🦞 by [Kent](https://github.com/tuanminhhole)
320
+
321
+ </div>