https://github.com/tillkuhn/travel-buddy
Demo project to plan travel destinations with Embabel
https://github.com/tillkuhn/travel-buddy
agentic ai embabel llm ramalama spring-ai spring-boot
Last synced: 5 days ago
JSON representation
Demo project to plan travel destinations with Embabel
- Host: GitHub
- URL: https://github.com/tillkuhn/travel-buddy
- Owner: tillkuhn
- Created: 2026-04-23T17:02:02.000Z (about 2 months ago)
- Default Branch: main
- Last Pushed: 2026-05-04T11:58:01.000Z (about 1 month ago)
- Last Synced: 2026-05-04T12:36:31.323Z (about 1 month ago)
- Topics: agentic, ai, embabel, llm, ramalama, spring-ai, spring-boot
- Language: HTML
- Homepage:
- Size: 2.06 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# πΊοΈ Embabel Travel Buddy
An [Embabel](https://embabel.com) agent demo in top of [spring-boot](https://spring.io/projects/spring-boot) that suggests travel destinations based on user preferences, running against a local LLM served by [RamaLama](https://github.com/containers/ramalama).
> [Embabel](https://github.com/embabel/embabel-agent) (Em-BAY-bel) is a framework for authoring agentic flows on the JVM that seamlessly mix LLM-prompted interactions with code and domain models. Supports intelligent path finding towards goals. Written in Kotlin but offers a natural usage model from Java. From the creator of Spring.
> [RamaLama](https://ramalama.ai/) is a LI for running AI models in containers on your machine. Instead of manually managing model runtimes and dependencies, you plug into an existing container engine such as Podman or Docker ([Source](https://www.redhat.com/en/blog/run-containerized-ai-models-locally-ramalama)).
## Prerequisites
- Java 25+
- Maven
- RamaLama running locally with a compatible model (e.g. `tinyllama`,`llama3.2`)
Install ramalama with brew
```bash
$ brew info ramalama
==> ramalama β: stable 0.19.0 (bottled)
Goal of RamaLama is to make working with AI boring
(...)
```
Run a particular model with ramalama
```bash
$ ramalama serve tinyllama --port 11434 --name llama --max-tokens=2000 --thinking=False
0.19: Pulling from ramalama/ramalama
Status: Image is up to date for quay.io/ramalama/ramalama:0.19
(...)
main: server is listening on http://0.0.0.0:11434
srv log_server_r: done request: POST /v1/chat/completions ::ffff:172.17.0.1 200
```
Example docker command invoked by ramalama behind the scenes
``` bash
docker run --rm --label ai.ramalama.model=ollama://library/granite3.1-dense:2b --label ai.ramalama.engine=docker \
--label ai.ramalama.runtime=llama.cpp --label ai.ramalama.port=11434 --label ai.ramalama.command=serve \
--security-opt=label=disable --cap-drop=all --security-opt=no-new-privileges --pull always -p 11434:11434 \
--label ai.ramalama --name ramalama --env=HOME=/tmp --init quay.io/ramalama/ramalama:0.20 llama-server Λ
--host :: --port 11434 --model /path/to/model --no-warmup --reasoning-budget 0 Λ
--alias library/granite3.1-dense --temp 0.8 --cache-reuse 256
```
## Does Model run on my hardware?
Check out [llmfit](https://github.com/AlexsJones/llmfit) ...
> A terminal tool that right-sizes LLM models to your system's RAM, CPU, and GPU. Detects your hardware, scores each model across quality, speed, fit, and context dimensions, and tells you which ones will actually run well on your machine.
```
brew info llmfit
==> llmfit β: stable 0.9.19 (bottled), HEAD
Find what models run on your hardware
```
**NOTE:** RamaLama defaults to the Ollama registry transport. To make it easier for users, RamaLama uses shortname files, which contain aliases for fully specified AI Models, on brew-ramalama it's located in `/opt/homebrew/Cellar/ramalama/0.20.0/libexec/share/ramalama/shortnames.conf`, or just run `ramalama info`
## Running the app
```bash
$ mvn spring-boot:run
```
```bash
Powered by Spring Boot 3.5.13
12:30:52.169 [main] INFO TravelApplication - Starting TravelApplication using Java 25.0.2 with PID 69558 2:30:52.170 [main] INFO TravelApplication - No active profile set, falling back to 1 default profile: "default"
12:30:52.498 [main] INFO OnMcpConnectionCondition - MCP connection condition NOT matched for [brave-search-mcp, fetch-mcp, wikipedia-mcp, docker-mcp] (...)
```
Then open [http://localhost:8080](http://localhost:8080) in your browser.
## What it does
Fill out the form with your travel preferences:
- **Region** β Americas, Southeast Asia, or Europe
- **Activities** β Hiking, Skiing, Diving, Beachtime, Culture, Shopping, Cycling (multi-select)
- **Additional Wishes** β any free-text notes (e.g. "family-friendly", "budget travel")
On submit, the app builds a prompt from your selections and invokes a single-step Embabel agent (`TravelPlannerAgent`) which asks the LLM for a concrete destination recommendation. The result is displayed on the next page.
## Example Agent & Service with Agent Invocation
```Java
// TravelPlannerAgent.java
@Agent(description = "Suggests a travel destination based on region, activities, and additional wishes")
public class TravelPlannerAgent {
@AchievesGoal(description = "Travel destination suggestion returned")
@Action
String suggestDestination(UserInput userInput, Ai ai) {
return ai
.withAutoLlm()
.generateText(userInput.getContent());
}
}
```
```Java
// TravelService.java
@Service
public class TravelService {
private final AgentPlatform agentPlatform;
public TravelService(AgentPlatform agentPlatform) {
this.agentPlatform = agentPlatform;
}
public TravelResult plan(TravelRequest request) {
String activitiesList = (request.activities() == null || request.activities().isEmpty())
? "no specific activities"
: String.join(", ", request.activities());
String prompt = """
You are an expert travel consultant. Based on the following traveler preferences, suggest ONE specific travel destination and explain in 3-5 sentences why it is a great fit.
If you don't know, say you don't know. Do not guess!
Region preference: %s
Desired activities: %s
Additional wishes: %s
Provide a concrete destination name and a compelling, personalized recommendation.
""".formatted(request.region(), activitiesList,
request.additionalWishes().isBlank() ? "none" : request.additionalWishes()).trim();
String suggestion = AgentInvocation
.create(agentPlatform, String.class)
.invoke(new UserInput(prompt));
List activities = request.activities() == null ? List.of() : request.activities();
return new TravelResult(suggestion, prompt, request.region(), activities, request.additionalWishes());
}
}
```
## Configuration
Key LLM settings in `src/main/resources/application.properties`:
| Property | Description |
|-------------------------------------------------|--------------------------------------------------------------------------------------------|
| `embabel.agent.platform.models.openai.base-url` | ramalama server URL, e.g. http://localhost:11434 |
| `embabel.models.default-llm` | Model name as reported by ramalama (`/v1/models`), e.g. gpt-4.1-mini |
| `embabel.agent.logging.personality` | Optional themed logging (`starwars`, `hitchhiker`, `montypython`, `severance`, `colossus`) |
To switch models, update `embabel.models.default-llm` in `application.properties` and add a matching entry in `src/main/resources/models/openai-models.yml`.
## Note on ramalama vs ollama
Despite the endpoint being `localhost:11434`, this app uses ramalama which speaks the **OpenAI wire protocol** β not the native Ollama protocol. The Embabel OpenAI starter (`embabel-agent-starter-openai`) is used accordingly.
## Web Resources
* [JCON2026 Talk "Comparing Agentic AI Frameworks for Java"](https://github.com/SandraAhlgrimm/ai-nutrition-planner/blob/main/slides.pdf) (slides) and [ai-nutrition-planner sample app](https://github.com/SandraAhlgrimm/ai-nutrition-planner) (sourcecode) and [Agent Deeplink](https://github.com/SandraAhlgrimm/ai-nutrition-planner/blob/main/embabel/src/main/java/com/example/nutritionplanner/NutritionPlannerAgent.java)
* [π€ Embabel Agent Examples
](https://github.com/embabel/embabel-agent-examples) for Kotlin and Java
* [Tripper: Embabel Travel Planner Agent](https://github.com/embabel/tripper) Sophisticated, realistic example application (from Embabel project)
* [Run containerized AI models locally with RamaLama](https://www.cedricclyburn.com/articles/ramalama-containerized-models) (good summary)
* [Build AI Agents in Java with Embabel: Step-by-Step Guide](https://bell-sw.com/blog/build-ai-agents-in-java-with-embabel-step-by-step-guide/) (good summary)
* [Baeldung: Creating an AI Agent in Java Using Embabel Agent Framework](https://www.baeldung.com/java-embabel-agent-framework) Brief coding Example