FAQ
Short answers to common questions about Burnbound. Each answer links to the page with the details.
What is Burnbound?
Spend control for AI agents. Before an agent refunds a Stripe charge, pays with a card from your Stripe Issuing account or pays an x402 or MPP API, Burnbound checks it against caps, team budgets, your own rules and approvals a person set. Burnbound decides; the money stays in your Stripe account or your wallet. Refunds and cards are in beta and start in Stripe's test mode. See Concepts.
What is x402?
x402 is an open protocol for paying for HTTP requests. A paid API answers 402 Payment Required with a price; the client signs a stablecoin payment and repeats the request with it; the seller settles the payment and serves the response. Burnbound pays x402 v2 with the exact scheme in USDC and, as an opt-in per agent, the MPP method tempo.charge in USDC.e on Tempo.
How do I limit what my AI agent can spend?
Give the agent a Burnbound policy and let it move money only through Burnbound. You set where it may send money, a maximum per payment, a daily cap and, if you want, a threshold above which a human must approve. The agent refunds with stripe_refund through your action gateway, pays x402 APIs with fetch_paid, and pays with its Stripe Issuing card, which Stripe checks with Burnbound before every authorization. Every payment outside those rules is denied before money moves. For API payments, the first payment to a receiving address your organization has not paid before also waits for a human, and so does a payment whose 402 reads like instructions to the agent. Start with the Quickstart; How to limit an AI agent's spending explains each limit.
Which agents and MCP clients are supported?
Any MCP client that can run a stdio server: Claude Code, Cursor and others. The server is the npm package @burnbound/mcp and runs with npx -y @burnbound/mcp@latest. The integrations show the setup for Claude Code, Cursor, Claude Desktop, the Vercel AI SDK, LangChain.js and Mastra. A JavaScript SDK exists but is not published yet; see SDK. Card payments need no MCP client at all: Stripe asks Burnbound directly.
Which networks and tokens can my agent pay with?
USDC over x402 on Base, or on Base Sepolia for tests. New agents pay on Base mainnet (eip155:8453). You can switch an agent to Base Sepolia (eip155:84532) when you create it in the onboarding, or later in its Network tab in the dashboard. You can also turn on MPP tempo.charge in USDC.e on Tempo (eip155:4217), or Tempo Moderato for tests; it is off by default and opt-in per agent. The policy denies any other network (network_not_allowed), asset (asset_not_allowed) or method (method_not_allowed). See Concepts.
Which currencies can refunds and cards use?
USD always works. Other currencies are converted to USD with the European Central Bank's daily reference rates, refreshed hourly and rounded up, so a payment never looks smaller to the caps than it is. That covers the about 30 currencies the ECB publishes, such as EUR, GBP, CHF and JPY. The server operator can pin a rate with FIAT_USD_RATES, which wins over the ECB's; any other currency without a pinned rate is refused with currency_not_supported and nothing moves. Caps and budgets are always in USD. See What counts as spend?.
Can a card payment wait for my approval?
No. Stripe gives Burnbound 2 seconds to answer a card authorization, and a card cannot wait for a person. A card payment that would need an approval is declined with approval_required, so set the agent's approval threshold above what its card should pay on its own. Refunds and API payments do wait for a human.
Can I approve from Slack?
Not in Slack itself, on purpose. Slack and email tell you that something is waiting, and their Approve and Reject buttons open it in the dashboard with that decision preselected. Approving moves money, so nothing is decided until you confirm in the dashboard, signed in as an owner or admin, where you see the amount, the payee, the reason and what the agent declared. The agent can never approve itself. See Human approval with Slack.
How do I test without real money?
Use Stripe's test mode for refunds and cards: a test charge to refund, and test authorizations on a test Issuing card. For paid APIs, set the agent's network to Base Sepolia, fund its wallet with test USDC from the Circle faucet, and pay an x402 API that charges on Base Sepolia. The flow is the same as on Base: caps, approvals and audit all apply. When you are done, switch the agent back to Base and fund the wallet with real USDC on Base. If the agent and the seller are on different networks, Burnbound denies the payment with network_not_allowed before anything is signed.
Do I need ETH in the wallet?
No. x402 payments are USDC transfer authorizations (EIP-3009) that the seller's facilitator submits on-chain, so the facilitator pays the gas. Your wallet only needs USDC. An MPP tempo.charge payment on Tempo needs no ETH either: the network fee, at most 0.01 USD, is paid in a stablecoin unless the seller sponsors it, so that wallet needs USDC.e on Tempo.
Does Burnbound hold my private keys or my money?
No. Burnbound never holds funds, card numbers, wallet private keys or your Stripe secret key. Refunds leave your Stripe account through an action gateway you host with your own restricted key, card money stays in your Stripe Issuing balance (Burnbound keeps only the webhook's signing secret, encrypted), and API payments go from your wallet to the seller. With Coinbase CDP, Burnbound stores your CDP API credentials encrypted so it can request signatures for the payments your policy allows. With client-side signing it knows only your wallet's address. See Security and trust model.
Which signing mode should I choose?
Choose Coinbase CDP if your agent can run shell commands or you need enforcement that does not depend on the agent's cooperation: the agent never holds anything that can sign. Choose client-side signing if you want the key on your own machine and accept that an agent with shell access could bypass the policy. See Concepts.
Can my agent raise its own limits?
No. An agent key can read its budget and payments, ask to authorize a payment and check its approvals, but it cannot widen its policy, approve a payment or create keys. Those actions need a human signed in to the dashboard. The only change an agent can make is a stricter one: when the user asks in the chat, propose_rule adds a rule that restricts the agent further and expires on its own.
How do I give several agents one shared budget?
Give each agent (or sub-agent) its own Burnbound agent and key, then, on the Team plan, put them in a subteam under Subteams in the dashboard and add a subteam rule, for example a monthly cap for the subteam. Each agent keeps its own caps, and together they never spend more than the subteam's cap, even when they pay at the same time from different machines. See How do subteams work?.
Can I add my own spending rules?
Yes. Besides the caps, each agent can have up to 50 custom rules (3 in all on the Free plan): deny some payments, hold them for approval, cap them per day, week or month, or only record them. Start from a template (weekly cap, cap per host, approval for a new payee) and try a rule in observe mode before you enforce it. See What are custom rules?.
What happens if Burnbound is down?
Your agent cannot pay, and nothing is charged. fetch_paid fails closed: if it cannot load the agent's allowed hosts or get a decision from Burnbound, it returns an error (control_plane_unavailable, api_unavailable, api_unreachable or api_timeout) and pays nothing. The action gateway refunds only on an allow, so without an answer nothing is refunded. For cards, Stripe applies what you chose in your Issuing settings for an authorization with no answer within 2 seconds: choose decline.
What happens if a payment does not complete?
fetch_paid returns paid only when the seller served the paid request (a 2xx). Anything else is a tool error with a code, and fetch_paid never signs the payment a second time on its own:
insufficient_funds: the paying wallet does not hold enough USDC on the payment's network, for exampleWallet 0x1234…5678 has 0.00 USDC on Base; this payment needs 0.01 USDC.Burnbound checks this before anything is signed, so nothing was paid. Add USDC to that wallet on that network and try again. The check is a courtesy, not a guarantee: the balance can change before the seller settles, and if Burnbound cannot reach the network it skips the check.payment_rejected_by_seller: the payment was signed and sent, but the seller refused it (often with another 402) and did not serve the response. The seller's reason, when it gives one, comes assellerReason, marked as untrusted text. A common cause is a wallet without enough USDC; another is a seller whose facilitator is failing. Check the cause before trying again.payment_not_completed: the seller answered the paid request with something else, such as a redirect or a server error. The payment may still settle; checklist_paymentsbefore retrying.payment_outcome_unknown: the seller's answer was lost. Checklist_paymentsto see whether the payment settled.
In the last three cases the error carries the intentId and the idempotencyKey, and Burnbound records the failed receipt with the seller's status and reason. Retrying fetch_paid with the same idempotencyKey never charges twice. See MCP tools.
To find out why a seller cannot be paid, without paying, the agent can call diagnose_payment, or you can run npx -y @burnbound/mcp@latest doctor <url> in a terminal, with no account. See MCP tools.
Does Burnbound take a fee on payments?
No. Burnbound does not take a percentage of what your agents spend: API payments go in full from your wallet to the seller, and refunds and card payments never pass through Burnbound. You pay for the plan, and no plan limits how much your agents spend.
| Free | Team | Business | |
|---|---|---|---|
| Price (VAT not included) | €0 | €39 per month or €390 per year | From €299 per month, billed yearly, on request |
| Managing people | 1 | 5 included, then €8 per month per person | Unlimited |
| Viewers (see spend and the audit log only) | 1 | Unlimited, free | Unlimited, free |
| Agents | 3 | Unlimited | Unlimited |
| Custom rules | 3 | Unlimited | Unlimited |
| Managed spend per month | No limit | No limit | No limit |
| Common fund, per-person caps, subteams, subteam and organization rules | No | Yes | Yes |
| Approval and low-balance notifications | Email and Slack | Email and Slack | |
| Audit history you can see | 30 days | 365 days | Unlimited |
Every plan has per-agent policies for refunds, cards and paid APIs, protection against prompt-injected payments, diagnose_payment, the catalog, the full MCP server and the approvals inbox. Team and Business add a CSV export and a monthly report per person and subteam (see How do I export spend and get a monthly report?). Business adds what larger companies ask for, built when a customer needs it: SSO, sending the audit to a SIEM, your own signer or HSM, an SLA and a DPA. The prices in force are on the pricing page, and Stripe handles billing.
How do the trial and the places work?
Each new organization gets 30 days of Team for free, with no card. We email you 7 days and 1 day before it ends.
- If you subscribe, you stay on Team: monthly or yearly, from the Plan page.
- If you don't subscribe, your organization moves to Free. It never moves to a paid plan unless you subscribe.
Team and Business take a limited number of paying organizations, and the pricing page shows how many places are left. If none are left, join the waiting list from Plan or from pricing: it doesn't commit you to pay and doesn't ask for a card, and we contact you when a place frees up. An organization that pays for Team keeps its place while it stays subscribed.
Moving to Free never loosens a limit or deletes anything. Sub-teams, the common fund and organization or sub-team rules keep applying wherever they restrict; on Free they can't be created or edited, but they can be deleted, and a person's daily sub-cap can be cleared but not set or changed. Agents and custom rules over the Free limits keep working. The whole audit log is kept: Free shows the last 30 days, and back on Team you see everything again. People over the Free limit become viewers, so nobody loses access to see.
The full conditions are in section 5 of the terms of service.
How do I stop an agent from paying?
Revoke its key in the dashboard. The MCP server then gets unauthorized on every call and cannot authorize payments. To stop it immediately without revoking the key, remove its allowed hosts or lower its daily cap: an empty host list allows no payment, card authorizations and refunds included. You can also turn off one method in its Network tab, for example Stripe Issuing · card, and keep the others.
Is there documentation for LLMs?
Yes. /llms.txt lists every documentation page with a one-line description, and /llms-full.txt has the whole documentation as one plain-text Markdown file.