API Keys Guide

How to get API keys for Claude, OpenAI, and Gemini, and what they cost.

Advisor Prep Hero is a "bring your own key" (BYOK) tool. You sign up for an account directly with one or more AI providers, get an API key, and paste it into Advisor Prep Hero. This means:

Your data stays private. AI requests go from your computer directly to the provider, Advisor Prep Hero is never in the loop.
You pay the provider directly. Advisor Prep Hero doesn't mark up AI usage. You pay whatever the provider charges, no middleman.
You control your usage. Set spending limits at the provider level. Switch providers or models any time.

You only need ONE provider to use Advisor Prep Hero. Most people start with Claude. You can add others later. Or skip the cloud providers entirely and run Ollama locally for free.

Don't have a key yet? Try the in-browser demo at advisorprephero.com/try first. It runs a real Advisor Prep Hero session against a sample workspace with a shared key, no signup, no credit card. Get a feel for the product before you sign up for any provider account.

      Where keys are stored: Advisor Prep Hero stores your API keys in your operating system's secure keychain, Keychain on Mac, Credential Manager on Windows, Secret Service on Linux. They're never written to a plain file, never logged, never sent anywhere except directly to the provider when you make a request. If you uninstall Advisor Prep Hero, your keys are removed from the keychain along with the app.
    

Anthropic (Claude)

Recommended for: long-form writing, document drafting, following complex instructions

Go to console.anthropic.com and sign up (free)
Add a payment method (you'll be billed only for actual usage)
Click API Keys in the sidebar
Click Create Key, give it a name like "Advisor Prep Hero", and copy the key (it starts with sk-ant-)
In Advisor Prep Hero, go to Settings → API Keys → Anthropic, paste the key, and click Save

Cost: Claude Sonnet (the default model) costs about $3 per million input tokens and $15 per million output tokens. A typical New Business Kickoff workflow run costs around $0.05-$0.20. Heavy daily use is usually $5-$15/month.

Free credits: Anthropic gives new accounts $5 in free credits. That's enough for ~50 workflow runs.

OpenAI (ChatGPT / GPT-4)

Recommended for: structured data extraction, JSON output, image-related tasks

Go to platform.openai.com and sign up (free)
Add a payment method
Click API Keys in the sidebar (or go to platform.openai.com/api-keys)
Click Create new secret key, give it a name like "Advisor Prep Hero", copy the key (it starts with sk-)
In Advisor Prep Hero, go to Settings → API Keys → OpenAI, paste the key, and click Save

Cost: GPT-4o costs about $2.50 per million input tokens and $10 per million output tokens. Comparable to Claude Sonnet for typical workflows. GPT-4o-mini is much cheaper (~$0.15 / $0.60) and fine for less demanding tasks.

Note: An OpenAI API account is separate from a ChatGPT Plus subscription. If you pay $20/mo for ChatGPT Plus, that does NOT give you API access, you need to set up billing separately on platform.openai.com.

Google AI (Gemini)

Recommended for: very long context windows, free tier for hobby use

Go to aistudio.google.com/app/apikey (sign in with a Google account)
Click Create API key
Copy the key
In Advisor Prep Hero, go to Settings → API Keys → Google, paste the key, and click Save

Cost: Gemini 2.0 Flash has a generous free tier (up to 15 requests/minute and 1500 requests/day at no charge). Beyond that, paid usage is roughly $0.075 per million input tokens, the cheapest of the three. Worth using as a "draft" model and switching to Claude or GPT for the final pass.

Ollama (local, free, no key needed)

Recommended for: total privacy, no cost per request, working offline

Ollama runs open-source models on your own machine. No API key, no provider account, no per-token charge. The tradeoff is hardware: you need a reasonably modern CPU and enough RAM (16 GB is comfortable for the smaller models, 32 GB+ if you want to run something like Llama 3.1 70B). On Apple Silicon it's genuinely fast.

Install Ollama from ollama.com (Mac, Windows, and Linux supported)
Pull a model from your terminal: ollama pull llama3.2 (or qwen2.5, mistral, etc.)
Make sure Ollama is running. By default it serves on http://localhost:11434
In Advisor Prep Hero, go to Settings → API Keys → Ollama. The "key" field is optional, leave it blank for the default localhost server, or paste a custom URL if Ollama runs on another machine on your network

Cost: Free. Your electric bill, basically.

Tradeoff: Local models are honestly really impressive in 2026, but they're still a notch below Claude Sonnet for long-form writing and complex instruction-following. Great for drafts, summaries, and anything you'd rather not send to a third party.

How to choose

If you want…	Use…
Best writing quality, long-form output	Claude (Sonnet or Opus)
Reliable structured output, JSON, code	GPT-4o
Cheapest, free tier for low-volume use	Gemini Flash
Multiple perspectives on the same prompt	All four (Advisor Prep Hero supports multi-model comparison)
Total privacy, offline use, no per-request cost	Ollama (local)

Setting spending limits

All three providers let you set hard spending caps on your account. We strongly recommend setting one, usage-based AI billing can get expensive fast if something runs in a loop.

Anthropic: Console → Plans & Billing → Cost limits
OpenAI: platform.openai.com → Settings → Limits → Set monthly budget
Google: Google Cloud Console → Billing → Budgets & alerts

A reasonable starting limit is $10/month per provider. Adjust up if you find yourself hitting it often.

What if my key gets compromised?

If you suspect your API key has been exposed (committed to a public git repo, shared in a screenshot, etc.):

Go to your provider's API keys page
Revoke the compromised key
Create a new key
Update Advisor Prep Hero with the new key (Settings → API Keys)

Compromised keys can be used to run up charges on your account, so revoke immediately if there's any doubt.

Why bring your own key instead of paying Advisor Prep Hero?

This is a deliberate design choice. By having you bring your own key:

Your privacy is real. If we routed AI calls through our servers, we'd be a man-in-the-middle who could (in theory) read your prompts. We don't want to be in that position. With BYOK, the cryptography goes user → provider, end of story.
No usage caps from us. You're limited only by your own provider account.
You don't pay markup. The provider's prices are what you pay.
You can use models we haven't enabled yet. When Anthropic releases a new Claude model, you can use it as soon as it's in their API, no waiting for us to update.

The downside is the small setup hurdle. Once it's done, it's done, you don't think about it again.