Billing & Credits

VibeFlow uses a usage-based billing model. You pay for the resources you use — projects, features, API calls, and optional add-ons.

Pricing Overview

Core VibeFlow Usage

API Calls

Add-Ons

Managing Your Subscription

Viewing Usage

1. Go to Billing in the sidebar
2. The Usage tab shows current consumption for each resource
3. Compare usage against your plan limits

Payment Methods

1. Go to Billing → Payment Methods
2. Add a credit card via Stripe
3. Set a default payment method
4. Remove old payment methods

Invoices

1. Go to Billing → Invoices
2. The first 20 invoices load by default, sorted newest first
3. Click Load more at the bottom of the table to fetch the next page
4. Each row links to the Stripe-hosted invoice page (View) and the PDF receipt (PDF)
5. Status badge (paid / open / void) is colored for at-a-glance scanning

Pagination uses Stripe's cursor model — the page you're looking at is fetched directly; the rest of your invoice history is only loaded when you click Load more. This keeps the page snappy even on accounts with hundreds of invoices.

📷 Screenshot needed: Invoices tab showing 20 paid rows + the "Load more" button below the table. Caption: "Invoice history paginates 20 at a time; click Load more for older invoices."

Per-Change vs. Cycle-End Invoicing

When VibeFlow's reconciler updates your Stripe subscription quantities (e.g., when your monthly resource counts change), the resulting proration is bundled into your next scheduled invoice rather than emitted as a standalone per-change invoice. You see one consolidated invoice per billing cycle instead of dozens of small per-event invoices.

User-initiated plan changes — upgrades, add-on purchases, cancellations — still produce an immediate invoice so you have a clear receipt for the action you took.

Changing Plans

1. Go to Billing → Subscription
2. View your current plan
3. Click Upgrade to see available plans
4. Select a new plan and confirm

Cancelling

1. Go to Billing → Subscription
2. Click Cancel Subscription
3. Provide a reason (helps us improve)
4. Confirm cancellation

Your access continues until the end of the current billing period.

Overage

If you exceed your plan limits:

• Overage charges apply at the per-unit rates listed above
• You'll receive a notification when approaching limits
• Overage is billed at the end of the billing cycle

Promo Codes

Apply a promotional code during plan upgrade to receive a discount:

1. Go to Billing → Subscription → Upgrade
2. In the checkout step, enter your promo code
3. The discount is applied to your invoice automatically

Promo codes may provide percentage discounts, fixed amount reductions, or free trial periods depending on the promotion.

Organization-Level Credit Enforcement

For Team and Enterprise plans, credits are tracked at the organization level:

• Org-level credit pool — All team members share a combined credit allocation from the plan tier
• Per-user usage tracking — Individual usage is tracked within the org pool
• Soft enforcement — When credits are exhausted, overage charges apply at per-unit rates (no hard cutoff for paid plans)
• Addon credits — Purchase additional credit packs that stack on top of the tier allocation

The Organization Usage tab in Billing shows a table view with columns for Product, Quantity Used, Limit, Overage, and Overage Rate.

LLM Gateway — Per-Class Token Billing

When the LLM Gateway forwards Anthropic /v1/messages requests, four token classes are tracked separately because each prices differently:

Service tier multipliers apply to the request total:

For high-cache-hit workloads (long system prompts, many turns), cache reads dominate the bill — and at the discounted 0.1× multiplier, the per-request cost can be substantially lower than the bare input_tokens × output_tokens multiplication you'd see in a naive ledger. The dashboards reflect the captured per-class breakdown, so you can see exactly where each dollar went.

📷 Screenshot needed: LLM Gateway billing dashboard showing a row with non-zero cache_creation_input_tokens and cache_read_input_tokens columns + the corresponding USD cost. Caption: "Per-class token breakdown — cache reads discount visible at the per-request level."

📷 Screenshot needed: Tier-grouped usage breakdown (standard / batch / priority) for an account that has used at least two tiers, so the difference in effective rate is visible. Caption: "Service tier multipliers shown side-by-side."

Note for Bifrost-routed models

Anthropic models routed via the OpenAI-compatible chat-completions endpoint (/v1/chat/completions) only expose cache reads through the upstream Bifrost SDK — cache_creation_input_tokens is collapsed into input_tokens by Bifrost's normalization layer and isn't recoverable through that path. Use the native Anthropic endpoint (/v1/messages) for the full per-class breakdown.

Enterprise / Contact Sales

For enterprise plans with custom pricing, volume discounts, or dedicated support:

• Visit the Contact Sales section in Billing
• Select your interests (Custom AI Agents, Vibe code to deploy enterprise design, etc.)
• A sales representative will contact you

If you're already on the Team tier and try to access an enterprise-only feature (e.g., LLM Gateway FinOps), the Upgrade button on the feature's nudge card is replaced with a disabled Enterprise button (with the tooltip "Contact sales to upgrade to the Enterprise tier."). Team is the highest self-serve tier — Enterprise is sales-contact, not a self-serve checkout.

📷 Screenshot needed: FinOps tab on the LLM Gateway page for a Team-tier user, showing the disabled "Enterprise" button (with tooltip) where the active "Upgrade" button used to be. Caption: "Team-tier users see a disabled Enterprise button on enterprise-only features — Enterprise upgrades go through sales, not self-serve checkout."