AI Setup

Ariv’s AI features are entirely optional. Every core feature — note editing, search, lifecycle buckets, tagging, the dashboard — works fully without any AI provider configured. AI is an enhancement layer you can turn on when you’re ready.

This guide walks you through setting up an AI provider so you can use features like Ask Brain, AI auto-tagging, and semantic search.

Step 1: Open AI Settings

Open Settings using Cmd+, (macOS) or Ctrl+, (Windows/Linux), then navigate to the AI & Intelligence section.

This is where all AI-related configuration lives.

Step 2: Choose a Provider

Set ai.provider to one of the following:

Provider	Key Required	Notes
Google Gemini (default)	Yes	Free tier available, great starting point
OpenAI	Yes	GPT-4o family
Anthropic (Claude)	Yes	Claude Sonnet and Haiku models
Ollama	No	Runs locally on your machine, fully offline

Tip

If you’re not sure which provider to pick, start with Google Gemini. It’s the default for a reason — Gemini 2.0 Flash is fast, capable, and Google offers a generous free tier so you can try AI features without spending anything.

For a detailed comparison of each provider, see the AI Providers guide.

Step 3: Enter Your API Key

Depending on your chosen provider, enter the corresponding API key:

Google Gemini

1. Go to aistudio.google.com and sign in with your Google account.
2. Create an API key.
3. Paste it into ai.gemini.apiKey in Ariv’s settings.

OpenAI

1. Go to platform.openai.com and sign in.
2. Navigate to API keys and create a new secret key.
3. Paste it into ai.openai.apiKey in Ariv’s settings.

Anthropic (Claude)

1. Go to console.anthropic.com and sign in.
2. Create an API key from the dashboard.
3. Paste it into ai.anthropic.apiKey in Ariv’s settings.

Ollama (Local)

No API key needed. Instead:

1. Install Ollama from ollama.ai if you haven’t already.
2. Set ai.ollama.url to your Ollama server address (default: http://localhost:11434).
3. Set ai.ollama.model to the model you want to use (e.g., llama3.1:8b).

Note

Ollama runs entirely on your machine. No data leaves your computer, and there are no API costs. However, it does require decent hardware — 8GB+ RAM is recommended for a smooth experience.

Step 4: Choose a Model

Set ai.model to the model you’d like Ariv to use. The available options depend on your provider:

Gemini Models

Model	Speed	Quality	Best for
Gemini 2.0 Flash (default)	Fast	Good	Everyday use, auto-tagging
Gemini 2.5 Flash	Moderate	Better	Balanced speed and quality
Gemini 2.5 Pro	Slower	Best	Complex Ask Brain queries

OpenAI Models

Model	Speed	Quality	Best for
GPT-4o	Moderate	Best	Most capable, complex questions
GPT-4o Mini	Fast	Good	Quick tasks, lower cost
o3-mini	Slower	Best	Reasoning-heavy queries

Anthropic Models

Model	Speed	Quality	Best for
Claude Sonnet 4.5	Moderate	Best	Balanced, high-quality output
Claude Haiku 4.5	Fast	Good	Quick tasks, lower cost

Custom Models

If your provider supports a model that isn’t in the dropdown, you can enter any model ID directly in ai.customModel. This overrides the dropdown selection and lets you use newer models as soon as your provider makes them available.

Recommended Starter Setup

For most users, the fastest path to working AI features is:

1. Provider: Google Gemini
2. API key: Free key from aistudio.google.com
3. Model: Gemini 2.0 Flash

This gives you a fast, capable AI setup with no upfront cost.

Usage Limits

Ariv includes usage limits on AI features to manage API costs:

Feature	Free Tier	Pro
Ask Brain queries	20/month	Unlimited
AI tag refinements	50/month	Unlimited
Semantic search	Unlimited (local)	Unlimited (local)

Note

Semantic search runs a local model on your machine and is always free regardless of your plan. It doesn’t count against any usage limit.

Ariv Pro ($7-8/month) removes all AI limits and gives you unlimited access to Ask Brain and AI-powered auto-tagging.

Verifying Your Setup

Once you’ve entered your API key and chosen a model, the easiest way to test that everything is working is to open a note and try Ask Brain — press Cmd+J (macOS) or Ctrl+J (Windows/Linux) and ask a question about your notes. If you get a response, your AI setup is good to go.

---

Related: AI Providers — detailed comparison of each provider | Auto-Tagging — how AI improves tag suggestions | Semantic Search — meaning-based note search