Agentic KG Services

Research Knowledge Graph & AI Paper Generation Platform

View the Project on GitHub djjay0131/agentic-kg

Service Inventory

Last Updated: 2026-04-17

This document provides a comprehensive inventory of all deployed services across the agentic-kg and Denario projects.

Quick Reference

Service Type Status Primary URL
agentic-kg-api-staging FastAPI ✅ Healthy https://agentic-kg-api-staging-tqpsba7pza-uc.a.run.app
agentic-kg-ui-staging Next.js ⚠️ Fixing https://agentic-kg-ui-staging-tqpsba7pza-uc.a.run.app
denario Streamlit ⚠️ Legacy https://denario-tqpsba7pza-uc.a.run.app
denario-app Streamlit ✅ Active https://denario-app-tqpsba7pza-uc.a.run.app

Agentic-KG Services

agentic-kg-api-staging

Type: FastAPI Backend Repository: djjay0131/agentic-kg Status: ✅ Healthy

URLs:

Purpose: REST API backend for the research problem knowledge graph system.

Features:

Health Check:

curl https://agentic-kg-api-staging-tqpsba7pza-uc.a.run.app/health
# {"status":"ok","version":"0.1.0","neo4j_connected":true}

API Documentation:

Environment Variables:

Dependencies:


agentic-kg-ui-staging

Type: Next.js Frontend Repository: djjay0131/agentic-kg Status: ⚠️ API URL configuration fix in progress

URLs:

Purpose: Web dashboard for exploring the knowledge graph and managing research problems.

Features:

Pages:

Build-Time Environment:

Dependencies:


Denario Services

denario (Legacy)

Type: Streamlit App Repository: djjay0131/Denario Status: ⚠️ Legacy - Consider deprecating

URLs:

Purpose: Original Denario Streamlit interface (older version).

Version: Commit 5d5713e Dockerfile: docker/Dockerfile.prod Install: pip install "denario[app]"

Environment Variables:

Issues:


denario-app (Active)

Type: Streamlit App Repository: djjay0131/DenarioApp Status: ✅ Active with improvements

URLs:

Purpose: Enhanced Denario Streamlit interface with production improvements.

Version: Commit b5e1270 Dockerfile: Dockerfile (root) Install: pip install denario_app

Improvements over legacy denario:

  1. File upload restrictions (.md and .txt only)
  2. Better error handling (UnicodeDecodeError)
  3. Vertex AI integration (GOOGLE_CLOUD_PROJECT, GOOGLE_CLOUD_LOCATION)
  4. Docker permission fixes
  5. Production warnings

Environment Variables:

Recent Commits:


Infrastructure

Neo4j Database

Type: Neo4j Graph Database Status: ✅ Running

Access:

Purpose: Graph database for storing research problems, papers, and relationships.

Credentials:

Terraform:


GCP Secret Manager

Secrets:

Access:

# List secrets
gcloud secrets list

# View secret value
gcloud secrets versions access latest --secret="OPENAI_API_KEY"

URL Formats

Cloud Run provides two URL formats for each service:

  1. Legacy Format (Project Number):
    • Pattern: https://{service}-{project-number}.{region}.run.app
    • Example: https://denario-542888988741.us-central1.run.app
    • Shown by: gcloud run services list
  2. New Format (Random String):
    • Pattern: https://{service}-{random}.{region-code}.a.run.app
    • Example: https://denario-tqpsba7pza-uc.a.run.app
    • Shown by: gcloud run services describe

Both URLs point to the same service and are fully functional.



Recommendations

Immediate Actions

  1. Fix agentic-kg-ui deployment - Rebuild with NEXT_PUBLIC_API_URL (in progress)
  2. Document denario vs denario-app - Clarify which to use
  3. Consider deprecating legacy denario service - Consolidate to denario-app

Future Improvements

  1. Set up custom domains for cleaner URLs
  2. Implement monitoring and alerting
  3. Create automated health check dashboard
  4. Document API endpoints with OpenAPI
  5. Set up automated deployment pipelines

Deployment Commands

List All Services

gcloud run services list --region=us-central1

Get Service Details

gcloud run services describe {service-name} --region=us-central1

Check Service Health

# agentic-kg API
curl https://agentic-kg-api-staging-tqpsba7pza-uc.a.run.app/health

# UI (check HTTP status)
curl -I https://agentic-kg-ui-staging-tqpsba7pza-uc.a.run.app

View Logs

gcloud run services logs read {service-name} --region=us-central1 --limit=50

Support

For issues or questions: