Auto start ollama service

README.md

@@ -18,32 +18,26 @@
 
 **OpenClaw-WebTop** runs OpenClaw securely with a need for a dedicate computer. When you are ready, you can move to your own machine docker environment if you would like.
 
+## Prequisite
+- Account with [Ollama](https://signin.ollama.com/) so that you can use an cloud based model for free with your daily free credits
+- Account with [NVIDIA Build](https://build.nvidia.com/) so that you can switch NVDIA free API as backup
 
 ## Quick start (TL;DR)
 
 1. Open this repository in a GitHub Codespace
-2. In the Codespace terminal run:
-
-```bash
-make start
-```
-
-3. Wait for the OpenClaw WebTop docker to boot. When the web desktop URL appears in the Codespace `Ports` Tab, click it to open the browser desktop.
-
-### Will be full automated
-
-4. In the webtop, open a terminal and run `ollama serve`
-5. Open another terminal and run `ollama signin`
-6. Sign in or sign up for a ollama free account.. and you get some cloud LLM token for free (reset every day and week)
-7. Download a cloud ollama model via `ollama pull kimi-k2.5:cloud`
+2. In the Codespace terminal run: `make start`. This will start a webtop OS that is accessible in the browser with ollama and openclaw pre-installed
+3. Wait for the OpenClaw WebTop docker to boot. When the web desktop URL appears in the Codespace `Ports` Tab, click it to open the webtop in a new browser.
+<img width="703" alt="launch-webtop-via-ports" src="https://raw.githubusercontent.com/gitricko/openclaw-webtop/gitricko-README.md/docs/launch-webtop-via-ports.png" />
+4. In the webtop, open a terminal, run `ollama signin` and sign in via the chrome browser
+<img width="703" alt="ollama-signin" src="https://raw.githubusercontent.com/gitricko/openclaw-webtop/gitricko-README.md/docs/ollama-signin.png" />
+7. Test connectivity via `ollama pull kimi-k2.5:cloud`
 8 Open another terminal and run `ollama launch openclaw --model kimi-k2.5:cloud`
 9. Follow the instructions till completion
 10. In the webtop, open a chromium browser and check if you can access this URI: http://localhost:127.0.0.1
 11. If it does not work: run `openclaw gateway run`
 12. When the URL is working, run `openclaw dashboard` it will give you to URI with a token at the end.
 
 
-
 <!-- ## Star History
 
 [![Star History Chart](https://api.star-history.com/svg?repos=gitricko/openclaw-webtop&type=date&legend=top-left)](https://www.star-history.com/#gitricko/openclaw-webtop&type=date&legend=top-left) -->

docker/Dockerfile

@@ -20,10 +20,10 @@ RUN ln -s /usr/local/lib/node_modules/npm/bin/npm-cli.js /usr/local/bin/npm
 # Copy the Ollama binary from the official image
 COPY --from=ollama-bin /usr/bin/ollama /usr/local/bin/ollama
 
-# (Optional) If you want Ollama to start automatically when the desktop loads,
-# you can add a custom init script to /custom-cont-init.d/
-# RUN echo 'ollama serve &' > /custom-cont-init.d/start-ollama.sh && \
-#    chmod +x /custom-cont-init.d/start-ollama.sh
+# Start Ollama to start automatically when the desktop loads,
+RUN mkdir -p /custom-cont-init.d
+RUN echo 'ollama serve &' > /custom-cont-init.d/start-ollama.sh && \
+   chmod +x /custom-cont-init.d/start-ollama.sh
 
 RUN npm install -g openclaw@${OPENCLAW_VERSION}