The Token Economy

What Are Genesis Tokens?

Genesis Tokens are the universal currency that powers everything in Genesis Writer — AI text generation, character portraits, character chat, and more. Instead of tracking usage across four different AI providers with four different pricing models, you have one unified token pool that works the same way regardless of which model you use.

Tokens Per Plan

Each subscription tier comes with a monthly token allowance. When your tokens run out, you'll need to wait for the next billing cycle or upgrade your plan.

Bonus Tokens

Genesis Writer gives you extra tokens in two situations:

• First-time subscriber bonus — when you subscribe to any paid plan for the first time, you receive a 25% bonus on top of your monthly tokens. This is a one-time perk per tier. If you subscribe to Writer and later upgrade to Creator, you get the bonus for each tier once.
• Referral discount — if someone signs up using your referral link, they get a 20% discount on their first subscription. See Referral Program for details.

For example, if you subscribe to the Writer plan ($19.99/mo, 3M tokens), your first month gives you 3M base tokens plus 750K bonus tokens, totaling 3.75M tokens.

What Costs Tokens?

Different actions cost different amounts of tokens. The cost depends on which model you use and how much output is generated.

Model Cost Tiers

Every model has a cost rating from 1 to 5 credits. This is a relative indicator — a 1-credit model is the cheapest, a 5-credit model is the most expensive. The actual token cost depends on how much context you send and how much text is generated, but the credit rating gives you a quick way to compare.

Image & Chat Costs

Beyond text generation, a few other features use tokens:

Image generation (portraits, illustrations, scene images) has a flat cost of 2,000 tokens per image regardless of style or content. Character chat costs approximately 1,000 tokens per message exchange.

How Token Reservation Works

Genesis Writer uses a reservation system to ensure you never get charged for text you don't keep. Here's the flow:

1. Reserve — when you click Write or Continue, tokens are reserved (held) from your balance based on the estimated maximum output.
2. Generate — the AI produces text while the reservation is active. The reservation has a brief expiry (a few minutes) as a safety net.
3. Finalize — when you click Keep, the actual cost is calculated from the real input and output token counts. This is usually less than the reservation. The difference is returned to your balance.
4. Cancel — if you click Discard or Stop, the entire reservation is refunded. Your balance returns to exactly what it was before.

The reservation lifecycle — tokens are only spent when you click Keep

Deduction Priority

When tokens are deducted, Genesis Writer follows a specific order:

1. Base tokens first — your monthly allocation (or one-time Starter grant) is used before anything else.
2. Bonus tokens second — the 25% first-time subscriber bonus is only consumed after your base tokens are depleted.

This matters because base tokens reset monthly while bonus tokens don't carry over. By spending base tokens first, the system ensures your bonus tokens act as a safety cushion that extends your capacity when you need it most.

Tracking Your Usage

You can monitor your token balance in several places:

• Settings page — your account settings show your current balance, usage history, and days until reset.
• Model selector — the model selector in the toolbar shows the credit cost for each model, so you can make informed choices before generating.

Track your token balance and usage from account settings

Low Token Warning

When your token balance drops below 20% of your plan's allocation, you'll see a warning indicator. This gives you time to either:

• Switch to cheaper models to stretch your remaining tokens further.
• Upgrade your plan for a larger monthly allocation.
• Wait for your monthly reset (paid plans refresh every 30 days).

Maximizing Your Tokens

A few practical strategies to get the most out of your token budget:

• Draft with cheap models, polish with premium ones. Use Grok 4.1 Fast or GPT-5 Mini (1 credit each) for rough drafts, then switch to Claude 4.5 Sonnet or Opus for your most important scenes.
• Use Continue instead of Write. Continue picks up from existing text, so the AI doesn't need to “figure out” what you want from scratch. This often produces better results with fewer retries.
• Provide good context. Fill in your project settings, character profiles, and locations. Better context leads to better first-try output, which means fewer discards and retries. See Context Building & Story Bible.
• Use writing styles. A well-configured writing style tells the AI exactly how to write, reducing the chance of getting output you don't like.
• Discard freely. Discarded generations are fully refunded. There's no penalty for trying a generation and not liking the result.

Next, learn about the fine-tuning knobs available in Advanced Settings.