See the filters' concepts for details.

By default, the filters can only be used from find endpoints generated by the Content Type Builder and the CLI. If you need to implement a filters system somewhere else, read the programmatic usage section.

Available operators

The available operators are separated in four different categories:


Filters are used as a suffix of a field name:

  • Not suffix or eq: Equals
  • ne: Not equals
  • lt: Lower than
  • gt: Greater than
  • lte: Lower than or equal to
  • gte: Greater than or equal to
  • in: Included in an array of values
  • nin: Isn't included in an array of values
  • contains: Contains
  • ncontains: Doesn't contain
  • containss: Contains case sensitive
  • ncontainss: Doesn't contain case sensitive


Find users having John as first name.

GET /users?firstName=John or GET /users?firstName_eq=John

Find products having a price equal or greater than 3.

GET /products?price_gte=3

Find multiple product with id 3, 6, 8

GET /products?id_in=3&id_in=6&id_in=8

Or clauses

If you use the same operator (except for in and nin) the values will be use to build an OR query

GET /posts?title_contains=article&title_contains=truc

Deep filtering

Find posts written by a user who belongs to a company with the name equal to Strapi GET /posts?


Querying your API with deep filters may cause performance issues. If one of your deep filtering queries is too slow, we recommend building a custom route with an optimized version of your query.

This feature doesn't allow you to filter nested models, e.g Find users and only return their posts older than yesterday.

To achieve this, there are two options:

  • Either build a custom route or modify your services
  • Use GraphQL


This feature isn't available for the upload plugin.


Sort according to a specific field.


Sort users by email.

  • ASC: GET /users?_sort=email:ASC
  • DESC: GET /users?_sort=email:DESC

Sorting on multiple fileds

  • GET /users?_sort=email:asc,createdAt:desc
  • GET /users?_sort=email:DESC,username:ASC


Limit the size of the returned results.


Limit the result length to 30.

GET /users?_limit=30

You can require the full data set by passing a limit equal to -1


Skip a specific number of entries (especially useful for pagination).


Get the second page of results.

GET /users?_start=10&_limit=10

Programmatic usage

You can use strapi-utils to parse the query params to Strapi's standared filters programmatically if you need it.

Extracting requests filters

To transform the query params to Strapi's standard filters a request, you can use the convertRestQueryParams function from strapi-utils.

const { convertRestQueryParams } = require('strapi-utils');

module.exports = {
  // when sending a request like GET /products?_sort:id&id=1
  fetchExpensiveProducts: (params, populate) => {
    const filters = convertRestQueryParams(params);

     *  filters = {
     *    start: 0,
     *    limit: 10,
     *    sort: [{ field: 'id', order: 'desc' }],
     *    where: [
     *      { field: 'id', operator: 'eq', value: 1 },
     *    ]
     * }

    // do sth with them

Querying data

We added a new API to query data base on the new filters API.

const { convertRestQueryParams, buildQuery } = require('strapi-utils');

module.exports = {
  find: async (ctx) => {
    // Convert params.
    const filter = convertRestQueryParams(ctx.request.query);

    return buildQuery({
      model: Article
      populate: []

SQL databases (strapi-hook-bookshelf)

If you are using a SQL database, calling buildQuery will return a Bookshelf Query on which you can call other functions (e.g count)

Mongo database

If you are using a mongo database calling buildQuery returns either a Mongoose Query or a custom query when used with deep filtering.

Custom Query

When using the deep filtering feature with mongo, we build an aggregation query to avoid too many round-trips with the mongo DB. Doing that means we don't get a Mongoose object as a response but instead a plain JS Object. This brings a some issues like no virtual fields available and no Mongoose lifecycles.

To deliver the best possible experience, we decided to rehydrate the Mongoose models, forcing us to override the Mongoose query

const query = buildQuery({
  model: Product, // you can use any models from strapi.models or strapi.plugins[pluginName].models
  filters: { limit: 10 },
  populate: [],

returns a query with the following functions

  • count => Returns an integer equal to the number of matching entities
  • lean => Returns the matching elements as Objects
  • then(onSucces, onFailure) => Calls the onSucces with an array of Mongoose objects.
  • catch(onError) => Promise catch
  • group(options) => Calls the aggregation group function of mongoose