How do Robo-Will credits work?
Every membership includes $3 in monthly credits, with the option to buy more. Learn how credits are used and why some requests cost more than others.
Usage pools
Every membership includes $3 in credits each month. Credits work like a balance — each message you send draws from it based on how many tokens are used.
There are two types of credits:
- Monthly credits — Included with your membership. These reset automatically at the start of each billing cycle.
- Purchased credits — Additional credits you can buy at any time from your AI usage dashboard. These roll over month to month and are spent before your monthly credits.
How tokens work
Every message you send to Robo-Will uses tokens. Tokens are small chunks of text — roughly 4 characters each. Every request has two sides:
- Input tokens — Your message, any selected page elements, and the conversation history
- Output tokens — Robo-Will's response
Both sides cost credits, and output tokens are typically 5–10× more expensive than input tokens. A typical request uses a few hundred to a few thousand tokens total.
Available models
Robo-Will supports a range of models from Anthropic, OpenAI, Google, and xAI. Each model has different strengths and costs. Here's a rough breakdown by tier:
Budget — great for quick questions
| Model | Good for |
|---|---|
| Claude Haiku 4.5 | Fast and affordable, good all-rounder |
| ChatGPT 5 Nano | Ultra-fast, simple lookups |
| Gemini 3 Flash | Fast, simple questions |
| Gemini 2.5 Flash Lite | Very cheap, basic tasks |
| Grok 4 Fast / Grok 4.1 Fast | Super affordable, huge 2M context |
Standard — balanced performance and cost
| Model | Good for |
|---|---|
| Claude Sonnet 4.6 | Coding, CSS, JavaScript — Will's pick |
| ChatGPT 5.2 | Writing, explanations |
| Gemini 3 Pro | Strong at coding and reasoning |
Premium — most capable
| Model | Good for |
|---|---|
| Claude Opus 4.6 | Complex, multi-step problems |
| ChatGPT 5.2 Codex | Complex, multi-step problems |
Why do some questions cost more?
A few things affect how much a request costs:
Model choice — This is the biggest factor. Premium models like Claude Opus can cost 5× more than a standard model like Claude Sonnet, and 25× more than a budget model like Gemini 3 Flash. The default model (Claude Sonnet 4.6) is a solid middle ground.
Response length — Longer, more detailed answers use more output tokens. A simple question like "how do I change a font color?" costs much less than asking Robo-Will to build out a full custom section.
Conversation history — Every message in a thread is re-sent as context. Long conversations gradually increase the input token count, even if your new message is short.
Tips for getting the most out of your credits
- Use the Select tool — Selecting a specific element gives Robo-Will precise context, leading to better answers with fewer back-and-forth messages.
- Keep conversations focused — Starting a new conversation for a new topic avoids carrying over tokens from previous exchanges.
- Use a budget model for simple tasks — If you just need a quick CSS tweak or a fast answer, Gemini 3 Flash or Claude Haiku will get the job done at a fraction of the cost.
- Save premium models for complex work — Switch to Claude Opus when you're working through something genuinely tricky.
FAQ
Do my monthly credits roll over? No — monthly included credits reset each billing cycle. Any credits you've purchased separately do roll over.
What happens when I run out of credits? You can purchase additional credits from your AI usage dashboard at any time.
Can I see how much each message cost? Yes — your usage dashboard shows a breakdown of credit usage over time.
Which model should I use by default? Claude Sonnet 4.6 is the default and the one I pick for most tasks. It's excellent at CSS, JavaScript, and Squarespace-specific problems. Switch to a budget model for quick questions and a premium model when you need it.