Pricing & Billing
What it costs, how billing works, and how to manage your balance.
Written By Philippe
Last updated 2 days ago
What You Pay For {#what-you-pay-for}
Every purchase on indie.money is for the output an agent produces when you run it — a finished result from the Builder's work product. You are not purchasing access to, and receive no rights in, any underlying model provider whose models the Builder chose to incorporate. No tokens, credentials, or model entitlements are transferred to you.
Pricing is set by each Builder and reflects the value of the result, not the cost of any individual input the agent consumes to produce it.
What It Costs {#cost}
Every run has up to two costs. They're separate and transparent.
Total Cost = Builder Fee + AI Costs
API Cost Note: When a Builder provides their own API keys, their pricing mode decides who pays for usage: buyers pay actual usage (default), usage included in the Builder's price, or usage plus the Builder's margin (25%--200%). When you bring your own keys, there is zero markup.
Example: Agent generates marketing copy.
- Builder: $0.10 + AI: $0.04 = $0.14 per run
- 100 runs = $14.00. Compare to a $50/month SaaS you'd barely use.
Choose higher-priced agents when your workflow needs heavier compute, because they get more compute per Code step from the default 5 seconds up to 300 seconds.
Your per-execution API budget {#api-budget}
Each run reserves a per-execution API budget — the maximum you will pay for AI provider calls during that single run. The Builder sets this when they publish the agent. Your wallet authorizes up to: builder fee + per-execution API budget.
If the run's actual API costs land below your budget, the unused portion is refunded automatically when the run settles (~5 min). If costs would exceed it, what happens depends on whose keys the agent uses — see "What happens when costs are high" below.
Three ways to pay {#payment-modes}
Each agent run is funded one of three ways. Producer or Builder chooses which mode is enabled per Agent; you (the buyer) pick one if more than one is available.
Add funds to your prepaid balance from the Balance tab in your dashboard. Withdraw anytime.
What happens when costs are high {#absorption}
Agents use one of three key sources for AI providers. Each has a different cost-overrun behavior:
The cost-overrun rule is set by the Builder when they publish the agent. Your run history shows which key source was used per call.
Why was my Agent paused? {#agent-paused}
An agent can pause for four reasons. Each shows a distinct banner in your dashboard. All are non-destructive — your purchase and scheduled runs are preserved.
Agent balance too low {#agent-paused-balance}
Your service budget ran out. The agent can't reserve the next run's maximum cost. Top up your service budget, then click Reactivate.
Per-execution API budget exceeded {#agent-paused-ceiling}
A single API call cost more than the budget reserved for that run. Common when the Builder uses their own keys and the AI provider returned a larger-than-expected response. You pay nothing for the unfinished run. Top up your service budget, then Reactivate.
Worked example. Bob runs an agent with a $0.01 per-execution API budget. A call costs $0.05 using the Builder's own key. The run stops, Agent pauses, Bob pays nothing for the failed run. Bob tops up and reactivates.
Platform API overage exhausted budget {#agent-paused-overage}
You used platform-provided keys, and recovered API costs from your service budget pushed it negative. The platform keys covered the call ceiling per-run, but your accumulated overage debits drained the budget. Top up to cover the outstanding overage, then Reactivate.
Manually paused {#agent-paused-manual}
You turned the agent off via the dashboard. No billing condition triggered this — it's a deliberate action. Click Reactivate to resume. Use this state to temporarily stop runs without losing the agent.
Worked example — Your prepaid balance. Charlie tops up his prepaid balance with $5 once. He runs his favorite weather agent 50 times that month at $0.10 per run — no signature popup each time, just instant runs. His prepaid balance drops to $0 after the 50th run; the Execute button shows "Top up to continue". He adds another $5 and keeps going.
How Billing Works {#billing}
TL;DR: We hold the max, you pay the actual, excess is refunded automatically when the run settles (~5 min). You never pay for uncompleted work.
The Cycle
What If a Run Fails?
Managing Your Balance {#balance}
You hold three separate balances, and topping up one does not fund another. Agent purchases and passes are paid in USDC directly from your wallet. Funding an agent balance credits that one agent's sponsored budget. Your prepaid balance — what your own runs draw from — is funded separately by depositing USDC to it, and you can withdraw it back to your wallet at any time for free (unspent run reservations are refunded automatically). The 10% conversion spread applies only when you cash Earnings out to USDC, not to prepaid deposits or refunds.
Where Your Balance Shows
Low Balance
When your balance is too low, the Execute button is disabled and a "Top Up" link appears. Purchase any agent or fund a budget with USDC to top up, then retry.
Cost Transparency {#breakdown}
Every run shows a complete breakdown in your run history.
Reading Your Receipt
If you brought your own API keys, those costs appear strikethrough — they're paid directly to the API provider, never billed through indie.money. The strikethrough shows what you saved.
If the agent called other agents (an agent-of-agents flow), expand Agent calls to see each sub-agent, the URL, and its individual breakdown.
Cost Components
Key promise: You never pay more than the authorized amount. Nothing is ever added on top of your authorized maximum.
Cost Protection: Every agent has a maximum API cost limit set by the Builder. If costs would exceed this limit, the run stops immediately. You only pay for completed work.
Platform API Overage & Funding Requirements {#overage}
When an agent uses platform-provided API keys (not your own), each run carries a potential overage charge. This is the maximum the platform might pay to the AI provider on your behalf for a single execution.
How It Works
Activation Gate
When you click Turn On, your service budget is checked against the funding requirement:
Why This Matters
Platform overage is charged per step, not per run. An agent with 3 platform-credentialed AI steps has 3× the potential overage of an agent with one. The funding requirement counts platform steps and multiplies by the highest per-call cap to guarantee the platform can cover worst-case costs.
Deactivation
If platform overage exhausts your service budget during a scheduled run, the agent is automatically deactivated with reason Platform API overage exhausted budget. Deposit funds and reactivate to resume.