A modern, Google search results-style product recommender that displays ML-powered recommendations with filtering, sorting, and AI explanations. Connects to the Ecosystem API for recommendations and user-interaction tracking.
# Create network (first time only)
docker network create ecosystem
# Run with defaults (uses https://meruntime1.ecosystem.ai)
docker run -d \
--name ecosystem-recommender \
--network ecosystem \
-p 54326:54326 \
-e ECOSYSTEM_API_URL=https://meruntime1.ecosystem.ai \
-e NUMBEROFFERS=10 \
ecosystemai/ecosystem-recommender:latest
Then open http://localhost:54326
| Variable | Default | Description |
|---|---|---|
ECOSYSTEM_API_URL | https://meruntime1.ecosystem.ai | Ecosystem API base URL |
NUMBEROFFERS | 10 | Number of product offers per request |
PORT | 54326 | Server port (container) |
LLM_API_KEY | — | Set for AI Explain feature (OpenAI or compatible API) |
LLM_MODEL | gpt-5 | LLM model name |
Using a local Ecosystem API?
ECOSYSTEM_API_URL=http://host.docker.internal:8091ECOSYSTEM_API_URL=http://ecosystem-runtime:8091docker network create ecosystemLLM_API_KEY for the AI Explain button (otherwise returns 503).Check available tags for your architecture (linux/amd64, linux/arm64).
See the project README for Docker Compose, full environment reference, and troubleshooting.
Content type
Image
Digest
sha256:6300d3ec1…
Size
133.2 MB
Last updated
18 days ago
docker pull ecosystemai/ecosystem-recommender