Billing & Wallet
Nexrall runs on a single wallet balance shared across every AI feature — Fullstack Agent Coding, Canvas, Nexrall Code, and the CLI.
How it works
- Every new account starts with a $5.00 signup bonus — no card required to try the product.
- Balance is only spent on AI usage: coding-agent builds and chat turns, Canvas sessions, and image generation.
- You always see your current balance — in the web app's account menu, and via
/balancein the CLI or the account menu in VS Code.
Monthly plans
You can subscribe to a monthly plan, or just top up when you need to — both feed the same wallet. A plan is a monthly allowance of spending power, granted to your wallet each cycle:
| Plan | Price | Monthly allowance |
|---|---|---|
| Basic | $20 / month | $20 of AI usage |
| Plus | $50 / month | $50 of AI usage |
| Pro | $200 / month | $200 of AI usage |
| Enterprise Team | Contact sales | Custom |
Plan billing is handled by Stripe Subscriptions and renews monthly until you cancel. Cancellation takes effect at the end of the current cycle — you keep whatever allowance you already received.
How plan allowance and top-up balance work together
- Plan allowance is spent first. While your current cycle has room, every AI charge is drawn from it.
- Once the allowance for the cycle is used up, further usage comes out of your permanent top-up balance.
- Unused plan allowance does not roll over. At renewal, any unspent allowance for the old cycle is removed and the new cycle's allowance is granted (so a $20 Basic plan grants exactly $20 each month, never more).
- Top-up balance never expires. It stays in your wallet until it is spent, regardless of which plan you're on.
Adding funds (top-up)
Open Settings → Billing in the web app and choose an amount to top up.
- Minimum top-up: $5.
- Credits your wallet 1:1 — a $50 top-up adds exactly $50 of spending power.
- Payment is handled by Stripe, embedded directly in the app (no redirect to an external checkout page for most payment methods).
What costs money
| Feature | What's billed |
|---|---|
| Fullstack Agent Coding | Model usage during a build/chat turn — scales with how much the agent reads, writes, and reasons about, not a flat per-app fee. |
| Canvas | Model usage while an artifact is being written or edited, plus AI image generation if you ask for one. |
| Nexrall Code / CLI | Model usage for agent turns, drawn from the same wallet. |
Models and cost
Nexrall is multi-provider, and different models cost different amounts per unit of usage. Anthropic's Claude models and OpenAI's GPT models are priced for the most capable reasoning, while DeepSeek and Qwen models trade some capability for a far lower per-token price — see Models for the catalogue and relative costs. Pick the default Claude Sonnet 5 for everyday work and reach for a cheaper or more capable model only when a task calls for it.
Pricing is usage-based and transparent: each charge is calculated from the provider's actual token cost plus a platform markup, and you can see the cost basis for every transaction in your account.
Insufficient balance
If your balance runs out mid-session, the current agent action is refused with a clear message rather than continuing and leaving you with a negative balance — top up (or wait for your next plan renewal) and pick up where you left off.
Refunds
Refunds and chargebacks on a top-up are handled automatically: if a payment is later refunded or disputed, the corresponding wallet credit is reversed. If you believe you were charged in error, contact support@nexrall.com.