PoC
Before Installation¶
Before installing KAI-Flow, make sure you have the following ready:
- Docker: KAI-Flow runs inside containers. You need Docker installed on your computer.
- Git: Required to download the project files.
- Environment Files: You will need to set up configuration files (
.env) for the backend and frontend. - Database: PostgreSQL is required. You can run it inside Docker or use an existing database.
Hardware Requirements¶
The requirements below are based on typical deployments. Your needs may vary depending on the number of users and the complexity of your workflows.
CPU Considerations¶
KAI-Flow does not require high CPU power for standard operations. Small cloud instances are sufficient for development and light usage. CPU usage may increase when running multiple workflows simultaneously or processing large documents. For production use with multiple concurrent users, we recommend 4 vCPUs or more.
Memory Considerations¶
KAI-Flow uses approximately 200-300MB of memory when idle. Memory usage increases based on your workflow complexity and the size of documents being processed. For production environments, we recommend at least 4 GB RAM, with 8-16 GB for better performance when working with large document collections.
Storage Considerations¶
Storage is needed for the database, uploaded files, and vector data. We recommend using SSD storage for better performance. Allocate at least 10 GB for basic usage and 20 GB or more for production deployments. Make sure to set up regular database backups and use persistent volumes in containerized environments.
Database Considerations¶
KAI-Flow uses PostgreSQL to store all application data including workflows, credentials, and execution history. The database must be created before starting the application, and the database user needs full table permissions. We recommend using a dedicated database for each KAI-Flow instance.
Quick Reference Table¶
| Component | Minimum | Recommended (Production) |
|---|---|---|
| CPU | 2 vCPUs | 4+ vCPUs |
| RAM | 4 GB | 8-16 GB |
| Storage | 10 GB SSD | 20 GB+ NVMe SSD |
| PostgreSQL | 13+ | 14+ |
Network Requirements¶
KAI-Flow runs on a microservices architecture. Ensure the following network configurations are set:
Internal Ports (Docker Network)¶
The services communicate internally via the kai_network bridge network:
- Backend: Internal communication on port
8000. - Frontend: Internal communication on port
3000. - Database: PostgreSQL default port
5432. - Kafka (Optional): Internal communication on port
9092.
External Access Ports¶
These ports must be exposed or allowed through your firewall:
- Backend API: Mapped to host port
${BACKEND_PORT:-23056}(default: 23056). - Frontend UI: Mapped to host port 23058.
- Widget: Mapped to host port 23059.
Outbound Connectivity¶
The backend requires outbound internet access to reach external AI providers and tool services:
api.openai.com(OpenAI API)api.tavily.com(Search Tool)api.smith.langchain.com(LangSmith Tracing)- External Kafka brokers (if you are connecting to remote Kafka clusters via KafkaTrigger node)