Attestify OS
Home·Builder Plans·Risk·Agent Controls
Agent Controls

Your agents shouldn't have your real card.

Give autonomous agents a programmable spending identity instead of a card — hard caps, pre-payment policy checks, and a durable audit trail. Every request is decided before the payment fires.

14-day free trial on Starter · No card required to try

The problem

Agents with cards are a liability.

Every agent you ship with access to a real card is an uncapped spend event waiting to happen — for you, your team, and your customers.

Solo developers

One runaway loop, one misread context, one API misconfiguration — your personal card takes the hit. Weekend tasks become Monday surprises.

🏢
Engineering teams

Shared agent credentials mean shared blast radius. Any agent that can reach the card can reach all of it. No per-task limits, no approval gates.

🔒
Enterprise shadow-IT

Finance can't audit what agents spent. Security can't trace which agent identity initiated a charge. Compliance is guesswork after the fact.

Real scenario

A research agent was given a corporate card to access data APIs during a vendor evaluation task. It hit a rate-limited endpoint, retried with exponential backoff — across 14 different API keys it found in the repo. $2,400 in weekend API charges. No pre-payment gate. No per-task cap. No audit trail that showed which agent identity initiated each call. Finance had a card statement. Security had nothing.

The difference

Card vs. Attestify

💳Agent with a card
  • Uncapped spend — the card limit is the blast radius
  • No per-task or per-agent isolation
  • Payment fires before any policy check
  • No agent identity in the audit trail
  • Finance gets a card statement; security gets nothing
  • Credential exposure = full account access
Agent with Attestify
  • Hard per-task and daily spend caps enforced before payment
  • Isolated wallet per agent or task scope
  • Every request evaluated by policy — payment only on pass
  • Agent identity, task ID, and policy rule in every audit record
  • Finance and security have exportable, structured receipts
  • Compromised credential = bounded blast radius
How it works

Pre-payment governance, not post-spend cleanup.

1
Agent requests payment
Your agent constructs a payment request with its task context — service class, amount, agent identity, and task scope.
2
Policy evaluation before spend
Attestify checks the request against your configured policy: per-task cap, daily cap, allowed service class. The payment does not fire during this check.
3
Approved or blocked — decision logged
If the request passes policy, the governed payment executes and a receipt is written. If it fails, the request is denied before spend and logged with the agent identity that initiated it.
4
Durable audit trail
Every decision — allowed or blocked — produces a structured audit record exportable to CSV or NDJSON. Finance and security have a trail, not a guess after the fact.
Live policy evaluation — example run
agent_idagent_procurement_04
requestmarket-data · $4.20
requestmarket-data · $3.10
requestanalytics-export · $18.00
policyper_task_cap: $12.00 — exceeded
receiptLogged
Implementation

Two config lines.

Replace a card credential with an Attestify key in your agent's environment config. Your agent code doesn't change — it's framework-agnostic, no SDK required — just the payment target and one header on every call.

agent-config.jsonBefore
{
  "payments": {
    "env": {
      "PAYMENT_METHOD": "card_4242_****",
      "SPEND_LIMIT": "none"
    }
  }
}
agent-config.jsonAfter
{
  "payments": {
    "base_url": "https://attestifyos.com/api",
    "env": {
      "ATTESTIFY_API_KEY": "your-tenant-key-here"
    }
  }
}

Send ATTESTIFY_API_KEY as the x-api-key header on any HTTPS request to POST /api/run — works from curl, fetch(), LangChain, CrewAI, AutoGen, LlamaIndex, or any HTTP client your agent already uses.

Start free. Scale as you grow.

14-day free trial on Starter. Your card is saved at signup but not charged until the trial ends. Cancel any time before day 14 and you will not be billed.

14-day free trial
Starter
Perfect for individuals and small teams getting started
$45/mo
  • Up to 2,500 governed agent runs / month
  • Full Router access
  • Signed receipts on every run
  • Verification & memory
  • Webhooks + dashboard
  • Email support
  • Quickstart onboarding guide
Most popular
Team
For teams that need more runs, more keys, and spend visibility
$95/mo

+ 1% of spend above $9,500

  • Up to 10,000 governed agent runs / month
  • Everything in Starter
  • Up to 5 tenant API keys
  • Per-tenant analytics
  • Spend rollups + daily breakdowns
  • Benchmark API
  • Priority email support
Enterprise
Compliance-grade controls for regulated organisations
Custom
  • Unlimited governed agent runs
  • Everything in Team
  • Unlimited tenant keys
  • Spend limits & budget enforcement
  • Approval workflows
  • Control Tower
  • Hash-chained audit ledger
  • SLA breach reporting & export
  • Priority Slack + dedicated support

Notes

  • Free trial — 14 days on Starter (monthly and annual). Your card is saved at signup but not charged until the trial ends. Cancel any time before day 14 and you will not be billed.
  • Team Monthly overage — 1% of spend above $9,500 in a billing period. If your team spends $12,000, you pay an additional $25 (1% × $2,500 excess). Overage does not apply to Team Annual.
  • Annual billing — charged upfront for the full year. Starter Annual: $456/yr ($38/mo effective). Team Annual: $948/yr ($79/mo effective).
  • Runs— a “run” is one governed agent call routed through the Attestify Router and recorded with a signed receipt.
  • Looking for Builder Plans or the Risk product line?
Secure checkout via Stripe · Cancel anytime · No setup fees · Questions? Email us

Frequently asked questions

Your agents are running. Are you governing them?

Switch from card to Attestify in two config lines. Hard spend caps and a cryptographic audit trail activate immediately.