Skip to main content
GET

Authorizations

api_key
string
query
required

Coin Metrics API key can be specified as ?api_key= query parameter.

Query Parameters

markets
string[]
required

Comma separated list of markets or market patterns like exchange-* or exchange-*-spot or *USDT-future.
Use a corresponding /catalog-all-v2/market-{dataType} endpoint for the full list of supported markets for a given data type.

type
enum<string>

Selects prediction-market order books. Without this parameter, prediction markets are excluded.

Available options:
event_contract
granularity
enum<string>
default:raw

Downsampling granularity of market order books and quotes. Supported values are raw, 1m, 1h, and 1d.

Available options:
raw,
1m,
1h,
1d
start_time
string

Start of the time interval.
This field refers to the time field in the response.
Multiple formats of ISO 8601 are supported: 2006-01-20T00:00:00Z, 2006-01-20T00:00:00.000Z, 2006-01-20T00:00:00.123456Z, 2006-01-20T00:00:00.123456789Z, 2006-01-20, 20060120.
Inclusive by default.
UTC timezone by default. Z suffix is optional and timezone parameter has a priority over it.
If start_time is omitted, response will include time series from the earliest time available.

end_time
string

End of the time interval.
This field refers to the time field in the response.
Multiple formats of ISO 8601 are supported: 2006-01-20T00:00:00Z, 2006-01-20T00:00:00.000Z, 2006-01-20T00:00:00.123456Z, 2006-01-20T00:00:00.123456789Z, 2006-01-20, 20060120.
Inclusive by default.
UTC timezone by default. Z suffix is optional and timezone parameter has a priority over it.
If end_time is omitted, response will include time series up to the latest time available.

start_inclusive
boolean
default:true

Inclusive or exclusive corresponding start_* parameters.

end_inclusive
boolean
default:true

Inclusive or exclusive corresponding end_* parameters.

timezone
string
default:UTC

Timezone name for start_time and end_time timestamps.
This parameter does not modify the output times, which are always UTC.
Format is defined by TZ database.

Example:

"America/New_York"

page_size
integer<int32>
default:100

Number of items per single page of results. Maximum value for unlimited order books is 100.
The value of this parameter is ignored if the endpoint supports the format parameter and its value is set to json_stream.

Required range: 1 <= x <= 10000
depth_limit
string
default:100

Book depth limit. Supported values are between 1 and 30000, 10pct_mid_price or full_book.

format
string
default:json

Format of the response. Supported values are json, json_stream.

paging_from
enum<string>
default:end

Where does the first page start, at the start of the interval or at the end.
The value of this parameter is ignored if the endpoint supports the format parameter and its value is set to json_stream.

Available options:
start,
end
limit_per_market
integer<int32>

How many entries per market result should contain. It is useful when multiple markets are requested.

pretty
boolean
default:false

Human-readable formatting of JSON responses.

next_page_token
string

Token for receiving the results from the next page of a query.
Should not be used directly. To iterate through pages just use next_page_url response field.

dataset
enum<string>
default:snapshots

Logical order book dataset to return.

snapshots (default) returns snapshot rows only.

updates returns Book Streams snapshot/update rows for migrated exchanges.

Caveats:

  • dataset=updates is currently supported for full_book and 30000 depths, raw granularity, json_stream format, paging_from=start, and migrated exchanges only.
  • When the updates dataset is requested, there are no guarantees on how snapshots are placed inside the data. Snapshots can appear at arbitrary timestamps and at irregular intervals.
  • Clients should reset their order book state on every snapshot row, not only on the first one. Snapshots may be emitted mid-stream, for example when the upstream feed is switched, and can carry book state that differs from prior updates. Treating every snapshot as a full state replacement is always safe while skipping snapshots is not.
  • When the updates dataset is requested, the response may contain redundant snapshots that do not change order book state and are there for state synchronization. Such redundant snapshots have the same timestamp and coin_metrics_id as the previous row. Clients may use this to detect redundant snapshots, but the safe default is still to reset state on every snapshot.
Available options:
snapshots,
updates
start_with_snapshot
boolean
default:false

Applies only when dataset=updates.

When false (default), rows are returned from the exact readable range. This is useful when the order book state is already initialized so the timestamp of the last row received can be used to resume downloading order book data.

When true, the response starts with a snapshot row when one is available at or before the effective dataset start. This is useful for order book state initialization. The API may include a snapshot from before start_time so client state can be initialized before the first update row is applied.

This parameter is ignored when dataset=snapshots.

Response

Time series of market order books.

data
object[]
required
next_page_token
string

Token of the next page results for a given request.

next_page_url
string

URL of the next page results for a given request.