Set the job once. Signed reports arrive on schedule.

Three steps.

1. Set the job once.Choose the tool, its input, a cadence, the callback URL and an end date. One call registers the job for your wallet and returns the exact schedule. Standing jobs run one read-and-report tool on your schedule.
2. The service runs it on schedule.At each run time Wever calls the tool with your input, through the same checks the tool applies to any call, and records the result.
3. Signed reports arrive at your callback.Each run posts a report signed by Wever that carries the tool’s own signed result. Every run stays on record for you to read later.

What a job can run.

Choose from 116 signed read-and-report tools in the current catalog. Each reads supplied material or supported sources and returns a signed result.

A standing job runs one read-and-report tool on your schedule and sends each signed report to your callback URL. Your agent reads the report and acts on it.

• Diligence checks and briefs. Refund windows, deadlines, subscriptions, quotes, policy fit, contract changes, claims, deliverables, invoices, terms, vendor evidence, insurance, audit evidence and agreements.
• Atlas readers. Issuer briefs, rights history, redemption, NAV evidence and published board evidence.
• Agent records. Endpoint reliability and honesty, directory status and listing checks, Bazaar declarations and wallet statements.
• Preflight checks. Identity, permissions, spending policies, approval chains, service terms, incident plans and continuity, using the inputs saved when you create the job.
• Briefs. Counterparty, wallet activity, fair value and job briefs.
• Lookups. Cross-venue market comparisons and escrow quotes.
• Watch evaluations. The state of your own Wever Watch watches, as a list or one watch in detail.

For an hourly endpoint check, choose wever_endpoint-reliability-check with your endpoint URL. Each run reads its availability, protocol behavior and response times, then delivers the signed report.

Inputs can carry {{run_date}} and {{run_at}}, so a daily refund check asks about today each time it runs. A job ends at the date you set, up to 30 days out, or after the number of runs you choose. You can cancel it whenever you like and future runs stop at once.

Digests and routing.

A digest gathers your watch fires and job results into one signed report per window: hourly, daily, weekly or your own interval. Quiet windows stay quiet unless you ask for empty digests too.

A routing rule sends a copy of matching results to a second address, such as a pager or a team inbox. Match on the tool that produced the result, a field of the report or both, for example every failed run or every watch fire above 120. The original delivery goes out exactly as before; routing adds one more.

A short worked example

Checking a refund window every day.

{
  "name": "wever_schedule-create",
  "arguments": {
    "tool": "wever_refund-eligibility-check",
    "input": {
      "purchase": {
        "item": "Wireless headphones",
        "amount": "129.99",
        "currency": "USD",
        "purchased_at": "2026-10-08",
        "delivered_at": "2026-10-10",
        "condition": "unused"
      },
      "policy": "You may return unused items within 30 days of delivery for a full refund.",
      "claim": {
        "requested_at": "{{run_date}}"
      }
    },
    "cadence": "daily",
    "start_at": "2026-11-05T13:00:00Z",
    "callback_url": "https://your-agent.example.com/hooks/runs",
    "expires_at": "2026-11-12T00:00:00Z"
  }
}

{
  "type": "wever.schedule-run.v1",
  "report": {
    "job_id": "job_…",
    "run_index": 6,
    "scheduled_for": "2026-11-10T13:00:00.000Z",
    "source_tool": "wever_refund-eligibility-check",
    "outcome": "succeeded",
    "summary": {
      "verdict": "outside_window"
    },
    "result": {
      "report": {
        "verdict": "outside_window",
        "…": "…"
      },
      "receipt": {
        "type": "wever.refund-eligibility-check.v1",
        "token": "eyJ…"
      }
    },
    "read_only": true,
    "action": "report_only"
  },
  "receipt": {
    "type": "wever.schedule-run.v1",
    "algorithm": "Ed25519",
    "token": "eyJ…"
  }
}

Call it from your agent.

Connect to https://weverlabs.com/mcp. Eight tools manage standing jobs:

• wever_schedule-create registers a job for your wallet: one read-only tool, its input, a cadence, a callback URL and an end date. It returns the job ID and the exact schedule Wever will run.
• wever_schedule-list lists your jobs with their state and next run, along with your digests and routing rules.
• wever_schedule-status shows one job in detail: runs completed, the last result, any failures, the next run and a full signed run report.
• wever_schedule-cancel ends a job. Future runs stop and past run records stay readable.
• wever_digest-create sends one signed digest per window with your watch fires and job results from that window.
• wever_digest-status shows a digest’s windows, deliveries and the last digest.
• wever_digest-cancel ends a digest.
• wever_alert-route posts a copy of matching watch fires and job results to one more callback URL.

Each call costs 0.10 USDC on Base, and a create call covers every run, digest or routed copy until it ends. Eligible calls share 10 free calls per wallet over a rolling 30 days with the other paid tools. Every call carries X-Wever-Wallet-Proof signed over the exact JSON arguments and the tool route, which is how Wever knows the job is yours. A paid call adds the x402 payment from the same wallet.

Reports arrive as a JSON POST with an Idempotency-Key. The contract lists the schedulable tools, cadence rules, delivery format, limits and the public verification key.