SilkClaw — Service Marketplace for AI Agents

Built for three audiences

A marketplace only works when every side wins.

🤖

AI Agents

Gain capabilities you don't have natively. Need weather data? DNS records? A vulnerability scan? Buy it in one call.

JSON in, JSON out — no parsing HTML
Micropayment pricing — most calls under a millicent
Input schemas returned on error for self-correction
Budget-aware — check balance before spending

🧑‍💻

People Using Agents

Give your agent a budget and let it work. Every purchase is logged. You stay in control.

Set a credit ceiling — agent can't overspend
Full transaction audit trail
Review what was purchased and what was delivered
One API key per agent — revoke anytime

🏪

Merchants

Wrap your API in a service definition and start earning from agent traffic immediately.

Define input schema, pricing, and delivery format
Agents discover your service automatically
Get paid per call — no minimums, no contracts
Payouts via Stripe Connect

One plugin, unlimited reach

An OpenClaw agent with the SilkClaw plugin can purchase any service in the marketplace — no extra integration needed.

How the agent sees it

The agent has a silkclaw tool. When it needs data it doesn't have — live weather, DNS records, a vulnerability scan — it calls the tool. SilkClaw handles discovery, payment, and delivery behind the scenes.

The agent doesn't need to know about individual APIs, auth flows, or rate limits. It just describes what it needs, and SilkClaw returns structured data.

What the human sees

A transaction log showing exactly what the agent bought, how much it spent, and what it got back. Full transparency, full control.

// Agent's tool definition
{
  "name": "silkclaw",
  "description": "Purchase data and compute services",
  "actions": [
    "search"  — find services by need
    "buy"     — purchase and receive delivery
    "balance" — check remaining credits
  ]
}

// Agent thinks: "I need to check if this
// dependency has vulnerabilities"

// Agent calls:
silkclaw.buy("svc_vuln_scan", {
  ecosystem: "npm",
  package: "lodash",
  version: "4.17.20"
})

// Agent receives:
{
  "vulnerabilities": [{
    "id": "CVE-2021-23337",
    "severity": "high",
    "fixed_in": "4.17.21"
  }]
}
        

Service catalog

Services launching with early access. Pricing in microcents — most calls cost a fraction of a penny.

How it works

From discovery to delivery in a single request.

1

Agent discovers services

The agent searches the catalog by capability, category, or keyword. No auth required to browse.

2

Agent checks the schema

Each service publishes its input schema, expected output format, pricing, and latency. The agent knows exactly what to send.

3

Agent buys the service

A single POST with the API key and input data. Credits are debited instantly. If inputs are wrong, the error includes the schema for self-correction.

4

Data delivered inline

The response contains the transaction receipt and the delivered data product. No callbacks, no polling, no async.

5

Human reviews the trail

Every purchase is logged with timestamp, service ID, cost, and input. The human can audit, set budgets, or revoke keys at any time.

Get early access

We're onboarding the first agents and merchants now. Join the waitlist to get your API key.

Your agent can buy what it needs