AI Setup

OctoFlow supports three AI providers out of the box. You bring your own API key (BYOK) — keys are stored locally in the macOS Keychain and never leave your Mac.

Supported Providers

Provider	Models	Key Required	Local
OpenAI	GPT-4o, GPT-4o-mini, GPT-4 Turbo	Yes	No
Anthropic	Claude 4 Sonnet, Claude 4 Haiku	Yes	No
Ollama	Llama 3, Mistral, Gemma, etc.	No	Yes

OpenAI Setup

Steps

1. Get an API key from platform.openai.com/api-keys
2. Open OctoFlow settings (menu bar icon > Settings)
3. Go to AI Provider tab
4. Select OpenAI and paste your API key
5. Choose a model (recommended: GPT-4o-mini for speed, GPT-4o for quality)

Anthropic Setup

1. Get an API key from console.anthropic.com
2. Open OctoFlow settings > AI Provider
3. Select Anthropic and paste your API key
4. Choose a model (recommended: Claude 4 Haiku for speed, Claude 4 Sonnet for quality)

Ollama (Local)

Run AI models entirely on your Mac — no API key needed, fully private.

1. Install Ollama: ollama.com/download
2. Pull a model:

   ollama pull llama3.2

3. Open OctoFlow settings > AI Provider
4. Select Ollama — OctoFlow auto-detects the local server at localhost:11434
5. Choose from your installed models

Tip

Ollama models run locally on your Mac’s GPU. Apple Silicon Macs with 16GB+ RAM handle 7B–13B parameter models well.

Changing the Default Model

You can set a default model for all AI actions, or override per-action in the toolbar:

1. Global default: Settings > AI Provider > Default Model
2. Per-action: Click the model name in the AI action result popover to switch

API Key Security

• Keys are stored in the macOS Keychain, encrypted at rest
• Keys are never transmitted anywhere except to the selected AI provider’s API
• No telemetry, no analytics, no OctoFlow servers involved