Defining a Schema

Your schema is the contract. Get it right and Runo extracts cleanly; get it wrong and you'll see nulls where you expect data.

Field Shape

{
  "field": "price",
  "type": "float",
  "example": 29.99,
  "hint": "Use sale price if present."
}

Key	Required	Notes
field	yes	The output key. Use the names your code expects. CamelCase is optimal.
type	yes	See the type table below.
example	yes	A representative value of the right type. Strongly recommended.
hint	no	Plain-English guidance for edge cases. Kept short (under ~140 chars).

Supported Types

Type	What It Accepts	Example Coercion
string	Anything; always a string	"Hello"
integer	Numbers, parsed from text	"35 years old" → 35
float	Numbers, parsed from text	"$1.2M" → 1200000.0
boolean	True/false signals	"✓ Verified" → true
date	ISO 8601 (YYYY-MM-DD); relative dates resolved to absolute	"yesterday" → "2026-05-07"
array<string>	JSON array of strings	["actress", "writer"]
array<integer>	JSON array of integers	[1, 2, 3]
array<float>	JSON array of floats	[1.5, 2.0]

Unresolvable fields return null.

Hints (When to Use Them)

Default behavior is usually fine. Reach for hint when:

• A page shows two values for the same concept and you want a specific one ("Use sale price if present.")
• The field name is ambiguous ("author" on a republished article - author of what?)
• The site uses non-obvious wording ("likes" is sometimes called "reactions")

Use them sparingly and only when necessary to avoid confusion.

Worked Examples

Product Page

[
  { "field": "title",   "type": "string",        "example": "MacBook Pro 14\"" },
  { "field": "price",   "type": "float",         "example": 1999.00, "hint": "Use sale price if present." },
  { "field": "currency","type": "string",        "example": "USD" },
  { "field": "inStock", "type": "boolean",       "example": true },
  { "field": "rating",  "type": "float",         "example": 4.6 },
  { "field": "tags",    "type": "array<string>", "example": ["laptop", "apple"] }
]

Article / Blog Post

[
  { "field": "headline",    "type": "string", "example": "OpenAI ships o3" },
  { "field": "author",      "type": "string", "example": "Cade Metz" },
  { "field": "publishedAt", "type": "date",   "example": "2024-12-20" },
  { "field": "summary",     "type": "string", "example": "A short summary of the article.", "hint": "1–3 sentences." }
]

Profile Page

[
  { "field": "firstName",  "type": "string",        "example": "Rachel" },
  { "field": "lastName",   "type": "string",        "example": "McAdams" },
  { "field": "age",        "type": "integer",       "example": 35 },
  { "field": "netWorth",   "type": "float",         "example": 12500000.00 },
  { "field": "isVerified", "type": "boolean",       "example": true },
  { "field": "tags",       "type": "array<string>", "example": ["actress", "producer"] }
]

Tips

• Keep schemas tight. 4–10 fields extract more accurately than 30. If you need more, split into two calls or a static key with a curated set.
• Use array<T> over delimiter strings. Don't ask for "tags" as a comma-joined string and split client-side. Declare array<string> and let Runo do it.
• Names matter. firstName and givenName produce subtly different extractions. Use the term your target sites tend to use.