# Mekler AI Swarm Session

> Mekler AI Swarm Session is a paid API for AI agents from api.mekler.ai, paid per call via x402, $5/call, status unknown (last checked 2026-10-30).

Starts a multi-worker AI swarm session to analyze, compare, or review supplied files using 1–32 cooperating AI workers with prepaid USDC billing.

## Facts

- Endpoint: POST https://api.mekler.ai/v1/swarm/sessions/start-session?utm_source=zero.xyz
- Price: $5/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-30
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/mekler-ai-swarm-session-8bb17e63
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_bWvEvgL3K_uxbW5Qe4raV

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability mekler-ai-swarm-session-8bb17e63 -d '<json body>'
```

Example prompt: I need to compare these 6 vendor proposals against our requirements doc — spin up 8 AI workers, give them up to 90 minutes, cap spending at $40, and tell me which vendors meet all the criteria and where each falls short.

## When to prefer this

Choose this endpoint when you need parallelized AI analysis across multiple files or large documents — especially for vendor comparison, document conflict detection, or dataset review — where isolated, verifiable computation matters and you want guaranteed correctness or a cost cap with unused-fund refunds. Prefer over single-model API calls when the task benefits from 2–32 cooperative workers and a structured brief with explicit acceptance criteria.

## Known failure modes

- Task failure: workers exhaust timeout or budget without meeting acceptance criteria — charged for used execution only
- Service failure: infrastructure or worker initialization failure — affected charges refunded
- Invalid brief: missing required fields (goal, acceptance_criteria, files list) — request rejected before payment
- Worker count not in allowed set (1,2,4,8,16,32) — schema validation error
- Budget too low or timeout out of range (must be 5400–7200s) — rejected
- Files exceed 200-entry limit or base64 encoding errors — rejected

## How this service works

Use 1–32 cooperating AI workers for supplied files: compare vendor proposals against requirements, review documents for conflicting claims, or analyze a supplied dataset. Prepay 5 USDC/worker; charge 1.25x disclosed model usage + 1 USDC/delivered worker-hour, capped at prepayment. Unused funds return after reconciliation. Task failures charge used execution; service failures refund affected charges. Isolated local tools; no external network, guaranteed correctness or speedup.

## Output

Returns a JSON object with a purchase_id and payment_state (e.g. 'pending'), representing the started swarm session. After execution, reconciled results include delivered output files meeting acceptance criteria, used worker-hours billed at 1.25x model cost plus $1/worker-hour, with unused prepaid funds returned.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "brief": {
   "type": "object",
   "required": [
    "goal",
    "constraints",
    "permitted_tools",
    "budget_micro_usd",
    "timeout_seconds",
    "expected_output"
   ],
   "properties": {
    "goal": {
     "type": "string",
     "maxLength": 20000,
     "minLength": 1
    },
    "constraints": {
     "type": "array",
     "items": {
      "type": "string",
      "maxLength": 4000
     },
     "maxItems": 50
    },
    "expected_output": {
     "type": "object",
     "required": [
      "files",
      "acceptance_criteria"
     ],
     "properties": {
      "files": {
       "type": "array",
       "items": {
        "type": "string",
        "maxLength": 240
       },
       "maxItems": 200,
       "minItems": 1
      },
      "acceptance_criteria": {
       "type": "array",
       "items": {
        "type": "string",
        "maxLength": 4000
       },
       "maxItems": 50,
       "minItems": 1
      }
     },
     "additionalProperties": false
    },
    "permitted_tools": {
     "type": "array",
     "items": {
      "enum": [
       "isolated_local_tools"
      ]
     },
     "maxItems": 1,
     "minItems": 1
    },
    "timeout_seconds": {
     "type": "integer",
     "maximum": 7200,
     "minimum": 5400
    },
    "budget_micro_usd": {
     "type": "integer",
     "maximum": 80000000,
     "minimum": 1
    }
   },
   "additionalProperties": false
  },
  "files": {
   "type": "object",
   "maxProperties": 200,
   "additionalProperties": {
    "type": "string",
    "format": "byte"
   }
  },
  "workers": {
   "enum": [
    1,
    2,
    4,
    8,
    16,
    32
   ],
   "type": "integer"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "purchase_id": "example",
  "payment_state": "pending"
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/mekler-ai-swarm-session-8bb17e63/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.mekler.ai](https://www.zero.xyz/host/api.mekler.ai/llms.txt)