Rate Limits

Every request consumes weight from your per-minute budget. Lighter queries cost less, heavier queries cost more.

Pricing tiers

Custom data type includes:

• Bitcoin Dominance
• Implied Volatility
• Skew

Weight formula

base_points = time_range_seconds / interval_seconds
base_weight = ceil(base_points / 1000) x cost_multiplier

When querying multiple exchanges with aggregation, an exchange multiplier applies:

exchange_multiplier = 1 + (0.2 x number_of_exchanges)

Depth-sensitive heatmap types also apply a maxDepth multiplier:

depth_multiplier = 1
if maxDepth > included_depth:
  depth_multiplier = 1 + (0.2 x floor((maxDepth - 1) / included_depth))
total_weight = ceil(base_weight x exchange_multiplier x depth_multiplier)

The minimum weight for any data request is 1. Metadata and discovery endpoints have zero weight.

Depth surcharge

For most data types, depth_multiplier stays at 1.0x. It only applies to:

If you omit maxDepth, or set it at or below the included depth, the depth multiplier stays at 1.0x.

Weight costs per data type

Polymarket analytics endpoints use a fixed weight instead of the point-based formula above:

Multi-exchange cost examples

Depth cost examples

These examples assume a 1,000-point request before exchange aggregation:

Burst capacity

Your tier has a burst capacity (2x your per-minute budget) that absorbs short traffic spikes. The bucket refills continuously at your sustained rate. For example, an Advanced tier user can briefly use up to 1,500 weight before being throttled.

Response headers

Check your usage

Use the rate limit endpoint to check your current quota without consuming any weight:

curl "https://api.kiyotaka.ai/v1/usage" \\
  -H "X-Kiyotaka-Key: YOUR_API_KEY"

import requests


response = requests.get(
    "https://api.kiyotaka.ai/v1/usage ",
    headers={"X-Kiyotaka-Key": "YOUR_API_KEY"}
)
data = response.json()

const response = await fetch(
  "https://api.kiyotaka.ai/v1/usage ",
  { headers: { "X-Kiyotaka-Key": "YOUR_API_KEY" } }
)
const data = await response.json()

{
  "limit": 750,
  "remaining": 680,
  "used": 70,
  "reset": 1712345678,
  "window_seconds": 60
}