Core Infrastructure

Every agent needs a second opinion.

One call. Your data verified, your reasoning checked, your action pre-flighted, your history remembered.

How It Works

Four checks. One call.

Manifest runs four independent analyses on every request and returns a unified verdict.

Verify

Is your data accurate?

Cross-references your data against independent sources. Catches stale prices, wrong numbers, and outdated metrics before you act on them.

Assess

Is your reasoning sound?

Checks if your premises are actually true, not just logical. Finds the gap between what you think is happening and what the data says.

Preflight

Will your action succeed?

Pre-flight checks for known risks, budget impact, and health status. Flags concentration risks, historical patterns, and execution hazards.

Remember

What happened last time?

Recalls past decisions on the same subject and tracks outcomes. Builds an unreplicable history that gets smarter with every check.

Not Just Crypto

Manifest verifies ANY data from ANY source.

Crypto, social, APIs, databases. If your agent consumed it, Manifest can check it.

Crypto

"SOL is $142"

Verified against 3 independent sources

"This account has 50K followers"

Re-fetched from source and confirmed

API

"Endpoint returned valid data"

Re-called and compared against schema

Live Example

See it in action

One natural language check. Four independent analyses. One verdict.

Request
POST /v1/manifest
// Headers
X-API-Key: cn-your-key
Content-Type: application/json

{
  "check": "SOL is at $142 with volume up 40%,
          I want to buy 100 SOL"
}

Response
{
  "verdict": "CAUTION",
  "confidence": 0.84,
  "verify": {
    "overall": "partial",
    "details": "Price confirmed ($142.48).
     Volume data is 2hr stale —
     current volume is DOWN 12%."
  },
  "assess": {
    "status": "FLAWED",
    "issue": "Premise 'volume up 40%'
     is based on stale data."
  },
  "preflight": {
    "status": "WARNING",
    "flags": [
      "Holder concentration 68%",
      "Similar setups preceded
       dumps 73% of the time"
    ]
  },
  "memory": {
    "note": "You've checked SOL 3 times
     today. Each time volume
     was disputed."
  },
  "summary": "Volume data is stale.
   Your premise is invalid.
   Recommend: WAIT."
}

Pricing

Pay per check. Scale to zero.

No subscriptions. No minimums. Credits deducted per call.

Quick

0.5 credits

Fast verification for time-sensitive decisions.

Verify + Preflight
Under 500ms
Cached source data

Standard

2.0 credits

Full analysis with all four checks.

All 4 checks
Full breakdown
Fresh source data
Memory context

Deep

5.0 credits

LLM-powered narrative with historical analysis.

All 4 checks + LLM narrative
Outcome statistics
Historical comparison
Confidence intervals

Identical requests within 5 minutes: FREE — served from memory. No credits deducted.

The Flywheel

It gets better the longer you use it.

Manifest compounds value over time. Your decision history becomes your competitive advantage.

Month 1

Your data verified against ClawNet's 390+ API endpoints. Instant cross-referencing from day one.

→

Month 3

Every new marketplace skill becomes another verification source. Coverage grows without you doing anything.

→

Month 6

"Agents who ignored CAUTION lost money 62% of the time." Outcome data starts validating the verdicts.

→

Month 12

Your decision history is unreplicable. Switching means losing everything. Manifest is your memory.