<div id="readability-page-1" class="page"><div>


    

    <article>
        <ul>
  <li><a href="#1-installation">1. Installation</a>
      - {:.} <a href="#model-provider">Model Provider</a>
      - {:.} <a href="#setup-openclaw-gateway">Setup OpenClaw Gateway</a>
      - {:.} <a href="#token-issue">Token Issue</a>
      - {:.} <a href="#slack">Slack</a></li>
</ul>

        <h2 id="1-installation">1. Installation</h2>

<div><pre><code><span>$ </span>git clone git@github.com:openclaw/openclaw.git 

<span># go to openclaw directory</span>
<span>$ </span>./docker-setup.sh
</code></pre></div>

<p>install brew manually.<br>
it seesm the docker container doesn’t have brew and during installation it fails if brew is not installed.</p>

<div><pre><code><span>mkdir</span> ~/.linuxbrew <span>&amp;&amp;</span> curl <span>-L</span> https://github.com/Homebrew/brew/tarball/master | <span>tar </span>xz <span>--strip</span> 1 <span>-C</span> ~/.linuxbrew

<span># set up environment variables </span>
<span>echo</span> <span>'export PATH="$HOME/.linuxbrew/bin:$HOME/.linuxbrew/sbin:$PATH"'</span> <span>&gt;&gt;</span> ~/.bashrc
<span>echo</span> <span>'export MANPATH="$HOME/.linuxbrew/share/man:$MANPATH"'</span> <span>&gt;&gt;</span> ~/.bashrc
<span>echo</span> <span>'export INFOPATH="$HOME/.linuxbrew/share/info:$INFOPATH"'</span> <span>&gt;&gt;</span> ~/.bashrc

<span># 3. 현재 세션에 적용</span>
<span>source</span> ~/.bashrc

</code></pre></div>

<h3 id="model-provider">Model Provider</h3>

<p><strong>vLLM</strong></p>

<p>you can serve local LLM like this.</p>

<div><pre><code>vllm serve openai/gpt-oss-20b <span>\</span>
    <span>--host</span> 127.0.0.1 <span>\</span>
    <span>--port</span> 8045 <span>\</span>
    <span>--tensor-parallel-size</span> 1 <span>\</span>
    <span>--gpu-memory-utilization</span> 0.3 <span>\</span>
    <span>--trust-remote-code</span> <span>\</span>
    <span>--async-scheduling</span> <span>\</span>
    <span>--max-num-batched-tokens</span> 8192 <span>\</span>
    <span>--max-model-len</span> 35096 <span>\</span>
    <span>--api-key</span> 1234
</code></pre></div>

<p>connect to Docker terminal and then change the connection information like this.<br>
no need to restart the openclaw container.</p>

<div><pre><code><span># change address to 172.17.0.1</span>
<span>sed</span> <span>-i</span> <span>'s/127.0.0.1/172.17.0.1/'</span> ~/.openclaw/openclaw.json

<span># change vLLM API key (1234)</span>
<span>sed</span> <span>-i</span> <span>'s/"VLLM_API_KEY"/"1234"/'</span> ~/.openclaw/openclaw.json
</code></pre></div>

<h3 id="setup-openclaw-gateway">Setup OpenClaw Gateway</h3>

<p>connect to Docker terminal and you can get a gateway token</p>

<div><pre><code><span>$ </span><span>echo</span> <span>$OPENCLAW_GATEWAY_TOKEN</span>
05c5df07ef01&lt;eradicated&gt;
</code></pre></div>

<ol>
  <li>http://127.0.0.1:18789/overview</li>
  <li>in Gateway Access, put the token in Gateway Token</li>
</ol>

<h3 id="token-issue">Token Issue</h3>

<p>If you happen to run <code>./docker-setup.sh</code> multiple times, it generates a new random gateway token. <br> 
this creates a mismatch between <code>.env</code> and <code>~/.openclaw/openclaw.json</code></p>

<div><pre><code><span>$ </span><span>grep </span>OPENCLAW_GATEWAY_TOKEN .env
<span>OPENCLAW_GATEWAY_TOKEN</span><span>=</span>05c5df0&lt;eradicated&gt;

<span>$ </span><span>grep</span> <span>-o</span> <span>'"token": "[^"]*"'</span> ~/.openclaw/openclaw.json
<span>"token"</span>: <span>"05c5df0&lt;eradicated&gt;"</span>
</code></pre></div>

<p>if both configurations are different, you need to update ~/.openclaw/openclaw.json file to match the .env token<br>
and then <strong>restart docker</strong></p>

<div><pre><code><span>$ </span>docker restart openclaw-openclaw-gateway-1
</code></pre></div>

<h3 id="slack">Slack</h3>

<div><pre><code><span>$ </span>./openclaw.mjs pairing list <span>--channel</span> slack
</code></pre></div>

    </article>

    
    

    
    
    
    
</div></div>

Incredible.AI

1. Installation
$ git clone git@github.com:openclaw/openclaw.git 
# go to openclaw directory
$ ./docker-setup.sh
install brew manually.
mkdir ~/.linuxbrew && curl -L https://github.com/Homebrew/brew/tarball/master | tar xz --strip 1 -C ~/.linuxbrew
# set up environment variables 
echo 'export PATH="$HOME/.linuxbrew/bin:$HOME/.linuxbrew/sbin:$PATH"' >> ~/.bashrc
echo 'export MANPATH="$HOME/.linuxbrew/share/man:$MANPATH"' >> ~/.bashrc
echo 'export INFOPATH="$HOME/.linuxbrew/share/info:$INFOPATH"' >> ~/.bashrc
# 3. 현재 세션에 적용
source ~/.bashrc
Model Provider
vLLM
you can serve local LLM like this.
vllm serve openai/gpt-oss-20b \
    --host 127.0.0.1 \
    --port 8045 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.3 \
    --trust-remote-code \
    --async-scheduling \
    --max-num-batched-tokens 8192 \
    --max-model-len 35096 \
    --api-key 1234
connect to Docker terminal and then change the connection information like this.
# change address to 172.17.0.1
sed -i 's/127.0.0.1/172.17.0.1/' ~/.openclaw/openclaw.json
# change vLLM API key (1234)
sed -i 's/"VLLM_API_KEY"/"1234"/' ~/.openclaw/openclaw.json
Setup OpenClaw Gateway
connect to Docker terminal and you can get a gateway token
$ echo $OPENCLAW_GATEWAY_TOKEN
05c5df07ef01<eradicated>
http://127.0.0.1:18789/overview
in Gateway Access, put the token in Gateway Token
Token Issue
If you happen to run ./docker-setup.sh multiple times, it generates a new random gateway token. 
 
this creates a mismatch between .env and ~/.openclaw/openclaw.json
$ grep OPENCLAW_GATEWAY_TOKEN .env
OPENCLAW_GATEWAY_TOKEN=05c5df0<eradicated>
$ grep -o '"token": "[^"]*"' ~/.openclaw/openclaw.json
"token": "05c5df0<eradicated>"
if both configurations are different, you need to update ~/.openclaw/openclaw.json file to match the .env token
restart docker
$ docker restart openclaw-openclaw-gateway-1
Slack
$ ./openclaw.mjs pairing list --channel slack

OpenClaw on Ubuntu + Docker Isolation

댓글