Skip to main content

Configuring Validata AI

This topic describes how to configure and launch Validata AI. Validata AI is an AI-powered assistant that runs locally and connects to either cloud-based AI providers or a local/remote self-hosted Ollama server for AI processing.

Prerequisites

Validata AI can be configured from the Validata UI. You can either use Get Started with Validata AI on the landing page, or select the Validata AI at the top right.

Before configuring Validata AI, ensure the following requirements are met:

  • Docker and Docker Compose installed on the server where Validata is running.

  • Inbound ports 9090 and 8001 must be open on the Validata server to allow access to the Validata AI interface and API.

  • Administrator-level access to the Validata server.

  • For cloud LLM configuration: an API key from your chosen provider (OpenAI, Google Gemini, or Anthropic Claude).

  • For Ollama configuration: access to an Ollama server with the gpt-oss:20b model installed.

  • For Ollama configuration: the Ollama server must meet the hardware requirements described in System requirements.

Note

Validata AI Docker containers run on the same machine as the Validata server. Remote deployment of Validata AI containers is not currently supported.

AI provider options

Validata AI supports multiple AI providers. Select the option that best meets your requirements for capability, cost, latency, and privacy:

Provider

Requirements

Notes

OpenAI

API key required

Cloud-based; requires outbound internet connectivity.

Google Gemini

API key required

Cloud-based; requires outbound internet connectivity.

Anthropic Claude

API key required

Cloud-based; requires outbound internet connectivity.

Ollama

Ollama server (local or remote)

Runs locally or on a remote server within your network. No API key required. All data processing remains within your infrastructure.

Note

To integrate with additional AI providers not listed above, contact Striim Support.

Configuring Validata AI

To configure Validata AI:

  1. Log in to Validata with administrator credentials.

  2. Select Get Started with Validata AI, or select Validata AI at the top right.

    Validata_Deploy_Admin_Roles_4.png

  3. Select your AI provider configuration option:

Configuring cloud LLM

To configure Validata AI with a cloud-based AI provider:

  1. Select Launch Validata AI using a cloud based LLM.

  2. From the Provider dropdown, select your AI provider (OpenAI, Google Gemini, or Anthropic Claude).

    Validata_ValidataAI_Working3.png

  3. Enter the Model Name for your selected provider. For example:

    Validata_ValidataAI_Working4_OpenAI.png

  4. Enter your API Key.

The following table shows example model names for each provider:

Provider

Example model name

OpenAI

gpt-4o

Google Gemini

gemini-2.5-flash

Anthropic Claude

claude-3-5-sonnet-20241022

Important

Handle API keys with the same confidentiality as passwords. Do not share them or commit them to version control systems. All API keys are stored encrypted using AES-256 encryption in Validata Historian.

Configuring Ollama

Validata AI supports two Ollama deployment options: local Ollama (running on the same machine as Validata) or remote Ollama (running on a separate server within your network).

Configuring local Ollama

To configure Validata AI with Ollama running locally:

  1. Select Launch Validata AI using self-hosted Ollama.

  2. Select Launch Ollama Locally.

Validata automatically sets the Ollama Server URL to http://host.docker.internal:11434.

Configuring remote Ollama

For advanced deployments, you can run Ollama on a dedicated remote server and configure Validata AI to connect to it.

To set up a remote Ollama server:

  1. Install Ollama on the server:

    curl -fsSL https://ollama.com/install.sh | sh

  2. Create or edit the Ollama systemd service file:

    sudo vim /etc/systemd/system/ollama.service

  3. Add the following content to the service file:

    [Unit]
Description=Ollama Service
After=network-online.target

[Service]
ExecStart=/usr/local/bin/ollama serve
User=ollama
Group=ollama
Restart=always
RestartSec=3
Environment="PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin"
Environment="OLLAMA_HOST=0.0.0.0:11434"

[Install]
WantedBy=multi-user.target

  4. Reload systemd and restart the Ollama service:

    sudo systemctl daemon-reload
sudo systemctl restart ollama

  5. Verify that the service is running:

    sudo systemctl status ollama

    Expected output:

    ● ollama.service - Ollama Service
     Loaded: loaded (/etc/systemd/system/ollama.service; enabled)
     Active: active (running)
     Main PID: 21034 (ollama)
     Memory: 10.8M
     CPU: 3.7s

    The logs should also indicate that Ollama is listening on port 11434:

    Listening on [::]:11434 (version 0.13.0)
discovering available GPUs...
starting runner

  6. Download the required model:

    ollama pull gpt-oss:20b

  7. Verify that Ollama is listening on port 11434:

    ss -antp | grep :11434

To configure Validata AI to use the remote Ollama server:

  1. Select Launch Validata AI using self-hosted Ollama.

  2. Select Use existing Ollama installation.

    Validata_ValidataAI_Working4_Ollama.png

  3. Enter the Ollama Server URL in the format http://<server-ip>:11434.

For example: http://192.168.1.100:11434

Warning

By default, Ollama doesn't include authentication. Striim recommends the following security practices for secure Ollama deployment:

  • Network isolation — Expose Ollama only on your internal network. Never expose it to the internet.

  • Firewall rules — Restrict access to specific IP addresses or subnets.

Launching Validata AI

After completing the configuration, select Launch AI Agent to start the service.

The system performs the following actions:

  • Validates the configuration parameters.

  • Downloads (if needed) and starts the required Docker containers.

  • Initializes the Validata AI Agent, and AI Store.

  • Makes Validata AI accessible in the Validata UI.

Once launched, you can access Validata AI from the Validata UI. See Validata AI for usage instructions.

Troubleshooting

This section describes common issues and their resolutions.

Configuration issues

Issue

Resolution

Invalid API key error

Re-enter the API key and verify that the key has the required permissions and quotas on the provider's platform.

Docker not operational

Confirm that Docker Engine and Docker Compose are installed and running on the Validata server.

Port conflicts

Ensure that ports 9090 (AI Server), 8001 (AI Agent), and 6333 (AI Store) are not in use by other processes.

Ollama-specific issues

Issue

Resolution

Model not found error

Run ollama pull gpt-oss:20b on the Ollama server to download the required model.

Connection refused error

Verify that the Ollama service is running and accessible at the configured URL. Check firewall rules and network connectivity.

Security considerations

Follow these security best practices when configuring Validata AI:

  • API keys are encrypted at rest using AES-256 encryption and are never displayed in plain text after initial entry.

  • Only users with administrator privileges can modify Validata AI configuration settings.

  • When using Ollama, all prompts and data remain within your organizational infrastructure.

  • When using cloud LLM providers, review the provider's Terms of Service regarding data privacy, data retention, and usage policies.

  • For remote Ollama deployments, restrict network access to trusted IP addresses and subnets.