Skip to main content
Filters use a single filter query parameter with an expression string:

​
Expression syntax

​
Operators

in_all and nin_all are available only when a repeated-record filter lists them in its discovered operators. They do not apply to scalar fields or every relationship filter. Use nin_any when no related record may match any of the requested values.

​
Examples

​
Filter by location

Locations are matched by numeric ID from the Dealroom location taxonomy, not by name. Look up an ID once via GET /reference/filters/location/values and reuse it:
Each value carries the two fields you need to identify it:
A name match is not unique - read type and parent_name to determine which ID you want for your search.
entity_count is null unless you pass include_counts=true. Counting the matching entities for every value is a separate scan per page, so it is opt-in. Add it only when you are rendering the counts, and prefer scoping the lookup with ?type= when you do - an untyped lookup for a broad term such as Europe has to count across every taxonomy level at once and can time out.
Narrow the lookup with ?type= to search a single taxonomy level - always worth doing, and required in practice if you also ask for counts. It accepts continent, country, state, city, and region:
Combined regions such as “London + Paris” are type: "region", alongside genuine geographic regions, so type=city rules them out. Five cities are still named “London” though, and only parent_name separates those.

​
Filter by industry (match any)

Industries (and other taxonomy values) live behind the taxonomy_id filter. Pass ?type=industry on the values endpoint to scope the lookup, then pipe-separate the resulting IDs in the filter expression:
The type parameter accepts: industry, sub_industry, sector, technology, business_model, income_stream, client_focus, sdg, ownership, techstack_category, growth_stage, investor_type.
A lookup scoped to the wrong type returns an empty 200, not an error — e.g. “Artificial Intelligence” is a technology tag, so ?q=artificial&type=industry finds nothing. When unsure, omit type to search across all tag types.

​
Combine multiple filters

Use and() to require all conditions:
Filter companies by the total size of the rounds their investors participated in:

​
Enum filters

Some filters accept a value from a fixed, closed set rather than free text or an ID. Discover the valid values with the values endpoint — same as id_lookup filters:
You can commit either the name (SERIES A) or its code (series_a) as the filter value. The following canonicalized enum filters match case-insensitively and reject unknown values with a 400 FILTER_VALIDATION_ERROR rather than silently returning no results: Because matching is case-insensitive, series a and SERIES A are equivalent:
A value outside the set returns an error instead of an empty result — see Errors:

​
Sorting

Use the sort parameter with a field name. Prefix with - for descending order.
The full list of accepted sort keys per resource is in the Filters & Sorting Reference.
A sort or filter key doesn’t always match a top-level response field. total_funding, for example, sorts and filters companies by the value returned at funding.total — there is no top-level total_funding field on the entity. Check the object pages from the API Reference overview for where a value appears in the response.

​
Available filters

For a complete list of all filters and sort keys grouped by scope, see the Filters & Sorting Reference. You can also fetch available filters programmatically:
The discovery response is the authoritative inventory for each public scope. It includes each filter’s type, supported operators, and any sub_types used for value lookup. Explicit cross-reference filters appear in their relevant scopes; dynamically constructible relationship paths are not listed exhaustively.

​
Further reading

Top fintech startups

Apply taxonomy lookup, filtering, and sorting in a complete query.

Pagination

Walk through a filtered result set without changing the query.