Configure Your LLM
Connect an AI model provider to power the Director
The Director needs an LLM provider to process messages and make tool-calling decisions. GNETiX supports multiple providers — you bring your own API key.
Add a provider
- Navigate to Org in the sidebar
- Select the LLM tab
- Click Add Provider
- Choose your provider and enter the required credentials
Anthropic (Claude)
The recommended provider for GNETiX. Supports extended thinking for complex reasoning tasks.
| Field | Value |
|---|---|
| Provider | Anthropic |
| API Key | Your Anthropic API key (sk-ant-...) |
OpenAI
| Field | Value |
|---|---|
| Provider | OpenAI |
| API Key | Your OpenAI API key (sk-...) |
Amazon Bedrock
Use Claude models through your AWS account. Data stays in your VPC.
| Field | Value |
|---|---|
| Provider | Amazon Bedrock |
| AWS Access Key ID | Your IAM access key |
| AWS Secret Access Key | Your IAM secret key |
| AWS Region | e.g. us-east-1 |
Create a dedicated IAM user with only bedrock:InvokeModel permissions for GNETiX. Never use root account credentials.
Azure OpenAI
Use GPT models through your Azure subscription.
| Field | Value |
|---|---|
| Provider | Azure OpenAI |
| API Key | Your Azure API key |
| Azure Endpoint | e.g. https://my-resource.openai.azure.com |
Model tiers
GNETiX uses three model tiers. The Director automatically selects the balanced tier for most interactions:
| Tier | Use case | Anthropic | OpenAI | Bedrock | Azure |
|---|---|---|---|---|---|
| Fast | Simple queries | Haiku | GPT-4o Mini | Haiku | GPT-4o Mini |
| Balanced | Most interactions | Sonnet | GPT-4o | Sonnet | GPT-4o |
| Powerful | Complex reasoning | Opus | GPT-4o | Sonnet | GPT-4o |
Extended thinking
When enabled, Claude shows its reasoning process before responding. This is visible in the Pipeline Monitor and helps understand how the Director decides which tools to call.
Enable it in Org → LLM → Extended Thinking toggle.
Extended thinking increases token usage significantly — each LLM iteration can use 5,000+ additional tokens for reasoning. Monitor your spend in the Org → Usage tab.
Budget caps
Set a monthly spend limit in Org → LLM → Monthly Budget. When the cap is reached, the Director stops processing new messages until the next billing cycle.
Leave the field blank to remove the cap.