AI Providers
Configure AI provider API keys and available models for your organization
FlowTrux supports multiple AI providers that power Agent nodes in your workflows. API keys are configured at the organization level and assigned to specific workspaces.
Supported Providers
Anthropic (Claude)
| Model | Best For |
|---|---|
claude-opus-4-6 | Most capable, complex multi-step reasoning, 1M context |
claude-sonnet-4-6 | Fast, high-quality, 1M context |
claude-opus-4-5 | Advanced reasoning and analysis |
claude-sonnet-4-5 | Balanced performance and speed |
claude-opus-4 | Strong reasoning, previous generation |
claude-sonnet-4 | Fast, reliable general-purpose tasks |
claude-haiku-4-5 | Lightweight tasks, high throughput, lowest cost |
OpenAI (GPT)
| Model | Best For |
|---|---|
gpt-5.4 | Most capable OpenAI model, 1M context |
gpt-5.3 | Agentic coding (Codex) |
gpt-5.2 | Advanced reasoning |
gpt-5 | Strong general-purpose performance |
gpt-5-mini | Cost-effective for simpler tasks |
o4-mini | Reasoning-optimized, compact |
Google (Gemini)
| Model | Best For |
|---|---|
gemini-2.5-pro | Most capable Gemini model |
gemini-2.5-flash | Fast responses, good balance |
gemini-2.0-flash | Lightweight, high-speed tasks |
Setup
1. Add an API Key
- Navigate to Settings in the sidebar (requires OWNER role).
- Open the AI Providers tab.
- Click Add Provider.
- Select the provider (Anthropic, OpenAI, or Google).
- Give it a name (e.g., "Production Anthropic Key").
- Paste your API key.
- Click Save.
The key is encrypted before being stored in the database. After saving, the key value is never exposed in the UI -- only the masked preview is shown.
2. Assign to a Workspace
API keys are not automatically available in all workspaces. You must explicitly assign them:
- Go to Settings > Workspaces.
- Click Edit on the target workspace.
- Under AI Providers, select which provider keys this workspace can use.
- Click Save.
Only models from assigned providers will appear in the Agent node model dropdown within that workspace.
How It Works at Runtime
When a workflow executes an Agent node:
- The engine looks up the selected model to determine which provider is needed.
- It loads the corresponding API key from the workspace's allowed providers.
- The key is decrypted in memory for the API call and never written to logs.
- If no matching provider is configured for the workspace, execution fails with a clear error message.
There is no fallback to environment variables. All AI provider configuration is done through the UI.
Multiple Keys
You can add multiple keys for the same provider. This is useful for:
- Separating environments (e.g., "Development" and "Production" keys)
- Cost tracking by assigning different keys to different workspaces
- Key rotation without downtime -- add the new key, reassign workspaces, then remove the old key
Plan Limits
Your organization's plan may limit the number of AI provider keys you can configure. Check Settings > Plan & Usage to see your current limits.