# Relaystation LLM Extract

> Relaystation LLM Extract is a paid API for AI agents from api.relaystation.ai, paid per call via x402, $0.01/call, status healthy (last checked 2026-09-15, last successful call 2026-08-30).

Pulls specified fields out of any input text and returns them as clean, schema-shaped JSON using a cost-efficient LLM.

## Facts

- Endpoint: POST https://api.relaystation.ai/v1/llm/extract
- Price: $0.01/call
- Payment: x402
- Status: healthy
- Last checked: 2026-09-15
- Last successful call: 2026-08-30
- Success rate: 100% of calls made through Zero
- Rating: 4.7 / 5 from 1 review
- Activations on Zero: 3
- Tags: x402
- Canonical page: https://www.zero.xyz/c/relaystation-llm-extract-3f700fa6
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_fnUNbBzuSyN4fNAQkdtpB

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 relaystation-llm-extract-3f700fa6 -d '<json body>'
```

Example prompt: Pull the sender name, date, and total amount out of this invoice text and give me the result as JSON: 'Invoice from Acme Corp, dated March 15 2025, total due $4,320.00 net 30.'

## When to prefer this

Choose this endpoint when you need to extract specific named fields from unstructured or semi-structured text and want clean JSON output without writing custom parsing logic. It is ideal for cost-sensitive pipelines where Opus-level accuracy is needed at sub-cent pricing. Prefer it over general-purpose LLM calls when the task is purely extraction (not generation or reasoning) and over regex/rule-based parsers when the text is too varied or natural-language to pattern-match reliably.

## Known failure modes

- Input text too ambiguous for field extraction — missing or null values returned for requested fields
- Schema or field list not understood — malformed or empty JSON output
- Text exceeds token limits — truncation or partial extraction
- x402 payment failure — 402 response if balance insufficient
- Rate limit exceeded — 429 response

## How this service works

Provider cost + 6%, itemized. Pull requested fields out of any text as clean JSON — schema-shaped output. ≈Opus-4.8 accuracy at ~1–5% of its cost (relaystation.ai/llm-validation). One of 11 vetted sub-cent LLM tasks on Relaystation — 155 tools, one balance.

## Output

Returns a clean JSON object containing the requested fields extracted from the input text, shaped to the schema or field list provided. Values are typed and labeled according to the requested output structure. Accuracy is comparable to Claude Opus 4.8 at approximately 1–5% of the cost.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "required": [
  "input"
 ],
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "type",
    "method",
    "bodyType",
    "body"
   ],
   "properties": {
    "body": {
     "type": "object",
     "properties": {
      "tier": {
       "enum": [
        "budget",
        "value",
        "best"
       ],
       "type": "string"
      },
      "input": {
       "type": "string",
       "minLength": 1,
       "description": "The text to process (DATA — never interpolated into the instruction)."
      },
      "params": {
       "type": "object",
       "required": [
        "schema"
       ],
       "properties": {
        "schema": {
         "type": "string",
         "minLength": 1
        }
       },
       "description": "Task-specific params for 'extract', validated against the task's param_schema."
      }
     }
    },
    "type": {
     "type": "string",
     "const": "http"
    },
    "method": {
     "enum": [
      "POST",
      "PUT",
      "PATCH"
     ],
     "type": "string"
    },
    "bodyType": {
     "enum": [
      "json",
      "form-data",
      "text"
     ],
     "type": "string"
    }
   },
   "additionalProperties": false
  },
  "output": {
   "type": "object",
   "required": [
    "type"
   ],
   "properties": {
    "type": {
     "type": "string"
    },
    "example": {
     "type": "object",
     "required": [
      "result",
      "truncated",
      "usage"
     ],
     "properties": {
      "usage": {
       "type": "object",
       "required": [
        "model",
        "inputTokens",
        "outputTokens",
        "providerCostMicros",
        "markupPct",
        "chargedMicros",
        "ceilingMicros"
       ],
       "properties": {
        "model": {
         "type": "string"
        },
        "markupPct": {
         "type": "number"
        },
        "inputTokens": {
         "type": "integer"
        },
        "outputTokens": {
         "type": "integer"
        },
        "ceilingMicros": {
         "type": "integer",
         "description": "The authorized ceiling the 402/hold reserved."
        },
        "chargedMicros": {
         "type": "integer",
         "description": "The actual charged amount (<= ceiling)."
        },
        "providerCostMicros": {
         "type": "integer"
        }
       }
      },
      "result": {
       "type": "string",
       "description": "The task output text (or JSON strin
… (truncated)
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/relaystation-llm-extract-3f700fa6/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.relaystation.ai](https://www.zero.xyz/host/api.relaystation.ai/llms.txt)