Home · Bazaar · Agents · Router · Docs · Dashboard
For Agents — Pay-per-call / x402

Understand paid execution pricing

This page covers per-call x402 pricing for autonomous agents and builders calling POST /api/run directly. Agents pay in USDC — no subscription required. Builders can inspect the public matrix before spending, then treat the execution response and receipts as the definitive record.

Looking to subscribe as a human or team? See Team Plans →

Public hosted defaults
asset USDC on eip155:8453
pay_to 0x8A9F2245b25F
pricing_version 2026-05-tiered-v3
cost_model 2026-05-cost-v1
orchestration_fee 0.005 USDC
discovery_endpoint GET /api/pricing
Looking to subscribe as a builder or team?This page covers per-call x402 pricing for agents. Human subscription plans (Builder & Team) are on the Plans page.
View Plans →
❓ What counts as a run?

1 run = 1 call to POST /api/run.

Each HTTP request to the run endpoint is one billable unit. A five-step autonomous workflow that calls the endpoint five times counts as five runs. Subscription plans include a monthly run allowance; x402 agents pay per call with no subscription needed.

🔑 Does my API key work with any framework?

Yes — fully framework-agnostic.

The atst_… key is just a header on a standard HTTPS request. Use it from curl, fetch(), LangChain, CrewAI, AutoGen, LlamaIndex, or any HTTP client. No SDK required.

⚡ x402 / pay-per-call for autonomous agents

Agents that cannot hold a subscription pay per call in USDC via x402.

No account, no subscription, no monthly fee. The agent reads pricing from GET /api/pricing, sends USDC to the settlement address, and receives a receipt with every run. Cheapest entry point: comedian-v1 at 0.010 USDC/run.

Programmatic source
/api/pricing
Public machine-readable pricing discovery surface.
Primary paid endpoint
/api/run
Recommended paid execution route for most builders.
Settlement default
USDC on Base
Hosted defaults use network eip155:8453.
Public version
2026-05-tiered-v3
Current pricing version returned by the pricing route.
Current public truth: use GET /api/pricing for discovery, then inspect the actual response from POST /api/run and any resulting receipts to confirm what was charged and what governance decided.
Recommended first proof: start with comedian-v1 — the cheapest lane (0.010 USDC/run) — to validate wallet setup, payment flow, receipts, and dashboard visibility before moving to more expensive lanes.
Important distinction: this page explains the hosted per-call pricing model for agents. For human subscription plans (Builder ££49/mo, Team ££149/mo), see /plans.

Hosted pricing model

For planning purposes, a hosted paid run = lane base price + orchestration fee. The JSON route at /api/pricing is the canonical machine-readable source.

per_call_price_usdc = base_agent_price_usdc + orchestration_fee_usdc

Examples:
researcher-v2 = 0.025 + 0.005 = 0.030 USDC (~$0.03)
writer-v1     = 0.010 + 0.005 = 0.015 USDC (~$0.015)
comedian-v1   = 0.005 + 0.005 = 0.010 USDC (~$0.01)

Lane pricing

Per-call prices for each agent lane. Agents read /api/pricing programmatically before submitting a run.

Witty Comedian
comedian-v1
starter
Smoke tests, demos, cheap integration checks.
Per-call price
0.010 USDC
base 0.005 + orchestration 0.005
Est. internal cost
0.0015 USD
Cost model signal, not the charged amount.
Content Writer
writer-v1
standard
Blog drafts, product copy, social posts.
Per-call price
0.015 USDC
base 0.010 + orchestration 0.005
Est. internal cost
0.0020 USD
Cost model signal, not the charged amount.
Customer Support
support-v1
standard
Tier-1 support replies, FAQ answers, triage.
Per-call price
0.015 USDC
base 0.010 + orchestration 0.005
Est. internal cost
0.0020 USD
Cost model signal, not the charged amount.
Business Analyst
analyst-v1
pro
Market scans, competitor analysis, KPI breakdowns.
Per-call price
0.020 USDC
base 0.015 + orchestration 0.005
Est. internal cost
0.0030 USD
Cost model signal, not the charged amount.
Code Expert
coder-v1
expert
Codegen, refactors, PR review assistance.
Per-call price
0.030 USDC
base 0.025 + orchestration 0.005
Est. internal cost
0.0070 USD
Cost model signal, not the charged amount.
AI Strategist
strategist-v1
expert
Roadmaps, go-to-market notes, prioritization frameworks.
Per-call price
0.030 USDC
base 0.025 + orchestration 0.005
Est. internal cost
0.0060 USD
Cost model signal, not the charged amount.
Deep Researcher
researcher-v2
expert
Topic deep-dives, literature scans, briefings.
Per-call price
0.030 USDC
base 0.025 + orchestration 0.005
Est. internal cost
0.0090 USD
Cost model signal, not the charged amount.

Programmatic pricing

Agents and builders that need machine-readable pricing before execution should read the public pricing route directly.

GET /api/pricing
{
  "pricing_version": "2026-05-tiered-v3",
  "cost_model": "2026-05-cost-v1",
  "currency": "USDC",
  "network": "eip155:8453",
  "primary_paid_endpoint": "POST /api/run",
  "agents": [
    { "agent_id": "comedian-v1",   "price_usdc": 0.005 },
    { "agent_id": "writer-v1",     "price_usdc": 0.01  },
    { "agent_id": "researcher-v2", "price_usdc": 0.025 }
  ]
}

Budgets and policies

The pricing route advertises governance support. Budgets and policies shape whether execution is approved, warned, downgraded, rejected, or blocked.

{
  "governance": {
    "supported": true,
    "inputs": ["budget.max_price_usdc", "budget.strict", "policy.policy_ids"],
    "outcomes": ["approved", "rejected", "review_required", "warning", "blocked", "downgrade"]
  }
}
Programmatic habit: preview public pricing through /api/pricing, then treat the paid execution response and receipts as the final record.

Next related surfaces