Skip to main content
The RWA API provides timeseries endpoints for tracking metrics over time and aggregating data across assets, tokens, networks, and asset classes.

Timeseries endpoints

Both endpoints use the same query language and response format.

The aggregate block

Timeseries queries use an aggregate block to control grouping, time resolution, and computation mode:

groupBy

Controls how results are grouped. Each unique group becomes a separate series in the response.

interval

The time resolution of each data point. quarter is supported on both timeseries endpoints, with either stock or flow mode.

mode

How values are computed within each interval. Defaults to flow on both timeseries endpoints. Example: For monthly supply, stock returns the latest available daily supply in each month, which need not be the last calendar day. For monthly mints, flow with sum adds the available daily mint values in each month. With interval: "day", mode has no effect because each point is already one day. For longer intervals, set mode: "stock" on state metrics. Otherwise the default flow mode with sum adds up the daily values and inflates the result.

aggregateFunction

Aggregation happens in two stages: first, aggregateFunction combines the matching entities into one value per date, measure, and group. Then the API combines those daily group values into intervals:
  • In flow mode, the same function combines the daily group values: sum, avg, min, or max. Flow is a period total only when the function is sum and the measure is additive.
  • In stock mode, the API selects the latest daily group value in the interval. aggregateFunction still controls how that day’s entities were combined; it does not combine different days.
For flow with avg, each returned daily group value has equal weight. This is a mean of daily means, not necessarily the mean of all underlying entity observations when the number of reporting entities varies by day.

Period labels and observation dates

For intervals longer than a day, the date in each [date, value] point labels the start of the calendar bucket, not the date of the selected stock observation or the end of flow coverage. Bucket starts are inclusive and the next bucket’s start is exclusive. These labels identify the period; they do not prove that every day in it has data or that your query includes the entire period.
Calendar buckets and their labels are always computed in UTC. Weeks start on Sunday, and months, quarters, and years start on the first day of the period. There is no query-level timezone setting. Treat returned YYYY-MM-DD labels as dates, not timestamps to convert into the client’s local timezone.

Relabeling a chart to period end

You can display January’s bucket at January 31, or the 2024 bucket at December 31, as a period-end display label. This does not change the data or establish that the value was observed on that date. In particular, do not present an unfinished month as a final month-end observation. For example, if January’s latest included daily point is January 18, its monthly label is still January 1. A chart may place it at January 31, but should mark it as partial and distinguish the display date from coverage through January 18. The aggregated [date, value] pair does not include a separate last-observation date or completeness flag. If you need the latest included daily date, request interval: "day" with the same filters and grouping, without groupLimit, and inspect that series. Even a daily snapshot date is not necessarily the original reporting date: some upstream measures carry values forward.

Date filters, unfinished periods, and gaps

Date filters apply to the underlying snapshot dates before interval aggregation. They do not filter the final bucket labels or expand a requested range to whole months, quarters, or years.
  • onOrAfter and onOrBefore include the specified date.
  • after and before exclude the specified date.
  • A mid-period start or end creates a partial bucket, even for a historical period. For example, January 10–20 data is still labeled January 1.
To include January 2026 without including February 1, add these clauses to an and filter alongside your measure and entity filters:
For an unfinished period, stock selects its latest included daily group value and flow combines the included daily group values so far. Neither mode waits for the period to close, prorates missing days, or forecasts the remainder. Data delays or filters can make the latest included date earlier than today. The interval aggregation itself does not create missing daily points, fill absent days with zero, or carry values across empty periods. A bucket with no daily points is omitted. Some upstream datasets already contain carried-forward values, so an available snapshot is not a guarantee of a fresh report; see Data quality. For grouped stock data, the API uses the latest date available for that group and measure, with only entities that have matching rows on that date. It does not independently select each entity’s latest date and combine stale and newer snapshots. Different groups can therefore use different observation dates even when their bucket labels match. If you set groupLimit, the synthetic “All Others” series combines series after interval aggregation. Its stock value can therefore combine observations from different dates within the same bucket.
An explicit null daily value is different from an absent date. Daily aggregation ignores null entity values, but returns null if all matching values that day are null. Stock selects the latest point even if its value is null; it does not search backward for a non-null value. If a daily group value is null, flow’s current numeric reduction treats it as zero for sum, min, and max, and includes it in the day count for avg. For example, daily group values [230, null] produce a flow average of 115. Inspect daily points if your application needs a different null policy.

Discovering available measures

Every timeseries query requires a measure_slug filter to specify which metric to retrieve. Use the meta endpoints to discover available measures:
The measure_slug field in the response lists all available measures as selectOptions. You can also browse measures in the Data Catalog. Common measures:

Response format

Timeseries responses differ from list endpoints. Each result contains a measure descriptor, a group descriptor, and an array of points:

Examples

Track an asset’s NAV over time

When no aggregate block is provided, the endpoint defaults to daily granularity grouped by the entity type.

Compare token supply across chains

Use groupBy: "token" to get a separate series for each on-chain deployment:
This returns one series per chain (e.g., “Ethereum USYC”, “Solana USYC”, “BNB Chain USYC”), so you can see how supply is distributed across networks.

Aggregate an entire asset class over time

Use groupBy: "asset_class" to sum across all tokens in a category. Query the token endpoint, because bridged_token_value_dollar returns no rows from the asset endpoint:
This gives you one data point per month for tokenized U.S. Treasuries, with Distributed and Represented tokens combined, so no tile on the site shows this total. To match the Treasuries page, add is_investable = true and tokenization_type_id = 3 (Distributed Value) or 2 (Represented Value). Do not sum net_asset_value_dollar this way: NAV is a per-unit value, so the sum adds up prices, not value.

Choosing the right endpoint