API reference contents

GET /2/tweets/search/stream

Streams Tweets in real-time based on a specific set of filter rules.

If you are using the academic research product track, you can connect up to two redundant connections to maximize your streaming up-time.

The Tweets returned by this endpoint count towards the Project-level Tweet cap.

Run in Postman ❯

Endpoint URL

https://api.twitter.com/2/tweets/search/stream

Authentication and rate limits

Authentication methods supported by this endpoint	OAuth 2.0 Bearer token
Rate limit	App rate limit: 50 requests per 15-minute window

Query parameters

Name	Type	Description
`backfill_minutes` Optional	integer	By passing this parameter, you can recover up to five minutes worth of data that you might have missed during a disconnection. The backfilled Tweets will automatically flow through a reconnected stream, with older Tweets generally being delivered before any newly matching Tweets. You must include a whole number between 1 and 5 as the value to this parameter. This feature will deliver all Tweets that matched your rules and were published during the timeframe selected, meaning that if you were disconnected for 90 seconds, and you requested two minutes of backfill, you will receive 30 seconds worth of duplicate Tweets. Due to this, you should make sure your system is tolerant of duplicate data. This feature is currently only available to the academic research product track.
`expansions` Optional	enum (`attachments.poll_ids`, `attachments.media_keys`, `author_id`, `entities.mentions.username`, `geo.place_id`, `in_reply_to_user_id`, `referenced_tweets.id`, `referenced_tweets.id.author_id`)	Expansions enable you to request additional data objects that relate to the originally returned Tweets. Submit a list of desired expansions in a comma-separated list without spaces. The ID that represents the expanded data object will be included directly in the Tweet data object, but the expanded object metadata will be returned within the `includes` response object, and will also include the ID so that you can match this data object to the original Tweet object. The following data objects can be expanded using this parameter: The Tweet author's user object The user object of the Tweet’s author that the original Tweet is responding to Any mentioned users’ object Any referenced Tweets’ author’s user object Attached media’s object Attached poll’s object Attached place’s object Any referenced Tweets’ object
`media.fields` Optional	enum (`duration_ms`, `height`, `media_key`, `preview_image_url`, `type`, `url`, `width`, `public_metrics`, `alt_text`)	This fields parameter enables you to select which specific media fields will deliver in each returned Tweet. Specify the desired fields in a comma-separated list without spaces between commas and fields. The Tweet will only return media fields if the Tweet contains media and if you've also included the `expansions=attachments.media_keys` query parameter in your request. While the media ID will be located in the Tweet object, you will find this ID and all additional media fields in the `includes` data object.
`place.fields` Optional	enum (`contained_within`, `country`, `country_code`, `full_name`, `geo`, `id`, `name`, `place_type`)	This fields parameter enables you to select which specific place fields will deliver in each returned Tweet. Specify the desired fields in a comma-separated list without spaces between commas and fields. The Tweet will only return place fields if the Tweet contains a place and if you've also included the `expansions=geo.place_id` query parameter in your request. While the place ID will be located in the Tweet object, you will find this ID and all additional place fields in the `includes` data object.
`poll.fields` Optional	enum (`duration_minutes`, `end_datetime`, `id`, `options`, `voting_status`)	This fields parameter enables you to select which specific poll fields will deliver in each returned Tweet. Specify the desired fields in a comma-separated list without spaces between commas and fields. The Tweet will only return poll fields if the Tweet contains a poll and if you've also included the `expansions=attachments.poll_ids` query parameter in your request. While the poll ID will be located in the Tweet object, you will find this ID and all additional poll fields in the `includes` data object.
`tweet.fields` Optional	enum (`attachments`, `author_id`, `context_annotations`, `conversation_id`, `created_at`, `entities`, `geo`, `id`, `in_reply_to_user_id`, `lang`, `public_metrics`, `possibly_sensitive`, `referenced_tweets`, `reply_settings`, `source`, `text`, `withheld`)	This fields parameter enables you to select which specific Tweet fields will deliver in each returned Tweet object. Specify the desired fields in a comma-separated list without spaces between commas and fields. You can also pass the `expansions=referenced_tweets.id` expansion to return the specified fields for both the original Tweet and any included referenced Tweets. The requested Tweet fields will display in both the original Tweet data object, as well as in the referenced Tweet expanded data object that will be located in the `includes` data object.
`user.fields` Optional	enum (`created_at`, `description`, `entities`, `id`, `location`, `name`, `pinned_tweet_id`, `profile_image_url`, `protected`, `public_metrics`, `url`, `username`, `verified`, `withheld`)	This fields parameter enables you to select which specific user fields will deliver in each returned Tweet. Specify the desired fields in a comma-separated list without spaces between commas and fields. While the user ID will be located in the original Tweet object, you will find this ID and all additional user fields in the `includes` data object. You must also pass one of the user expansions to return the desired user fields: `expansions=author_id` `expansions=entities.mentions.username` `expansions=in_reply_to_user_id` `expansions=referenced_tweets.id.author_id`

Example requests

      curl https://api.twitter.com/2/tweets/search/stream -H "Authorization: Bearer $BEARER_TOKEN"

      curl https://api.twitter.com/2/tweets/search/stream?backfill_minutes=5&tweet.fields=created_at%2Cconversation_id&expansions=author_id -H "Authorization: Bearer $BEARER_TOKEN"

Example responses

      {
  "data": {
    "id": "1067094924124872705",
    "text": "Just getting started with Twitter APIs? Find out what you need in order to build an app. Watch this video! https://t.co/Hg8nkfoizN"
  }
}

Response fields

Name	Type	Description
`id` Default	string	Unique identifier of this Tweet. This is returned as a string in order to avoid complications with languages and tools that cannot handle large integers.
`text` Default	string	The content of the Tweet. To return this field, add `tweet.fields=text` in the request's query parameter.
`created_at`	date (ISO 8601)	Creation time of the Tweet. To return this field, add `tweet.fields=created_at` in the request's query parameter.
`author_id`	string	Unique identifier of this user. This is returned as a string in order to avoid complications with languages and tools that cannot handle large integers. You can obtain the expanded object in `includes.users` by adding `expansions=author_id` in the request's query parameter.
`conversation_id`	string	The Tweet ID of the original Tweet of the conversation (which includes direct replies, replies of replies). To return this field, add `tweet.fields=conversation_id` in the request's query parameter.
`in_reply_to_user_id`	string	If this Tweet is a Reply, indicates the user ID of the parent Tweet's author. This is returned as a string in order to avoid complications with languages and tools that cannot handle large integers. You can obtain the expanded object in `includes.users` by adding `expansions=in_reply_to_user_id` in the request's query parameter.
`referenced_tweets`	array	A list of Tweets this Tweet refers to. For example, if the parent Tweet is a Retweet, a Retweet with comment (also known as Quoted Tweet) or a Reply, it will include the related Tweet referenced to by its parent. To return this field, add `tweet.fields=referenced_tweets` in the request's query parameter.
`referenced_tweets.type`	enum (`retweeted`, `quoted`, `replied_to`)	Indicates the type of relationship between this Tweet and the Tweet returned in the response: `retweeted` (this Tweet is a Retweet), `quoted` (a Retweet with comment, also known as Quoted Tweet), or `replied_to` (this Tweet is a reply).
`referenced_tweets.id`	string	The unique identifier of the referenced Tweet. You can obtain the expanded object in `includes.tweets` by adding `expansions=referenced_tweets.id` in the request's query parameter.
`attachments`	object	Specifies the type of attachments (if any) present in this Tweet. To return this field, add `tweet.fields=attachments` in the request's query parameter.
`attachments.media_keys`	array	List of unique identifiers of media attached to this Tweet. These identifiers use the same media key format as those returned by the Media Library. You can obtain the expanded object in `includes.media` by adding `expansions=attachments.media_keys` in the request's query parameter.
`attachments.poll_ids`	array	List of unique identifiers of polls present in the Tweets returned. These are returned as a string in order to avoid complications with languages and tools that cannot handle large integers. You can obtain the expanded object in `includes.polls` by adding `expansions=attachments.polls_ids` in the request's query parameter.
`geo`	object	Contains details about the location tagged by the user in this Tweet, if they specified one. To return this field, add `tweet.fields=geo` in the request's query parameter.
`geo.coordinates`	object	Contains details about the coordinates of the location tagged by the user in this Tweet, if they specified one. To return this field, add `tweet.fields=geo.coordinates` in the request's query parameter.
`geo.coordinates.type`	string	Describes the type of coordinate. The only value supported at present is `Point`.
`geo.coordinates.coordinates`	array	A pair of decimal values representing the precise location of the user (latitude, longitude). This value be `null` unless the user explicitly shared their precise location.
`geo.place_id`	string	The unique identifier of the place, if this is a point of interest tagged in the Tweet. You can obtain the expanded object in `includes.places` by adding `expansions=geo.place_id` in the request's query parameter.
`context_annotations`	array	Contains context annotations for the Tweet. To return this field, add `tweet.fields=context_annotations` in the request's query parameter.
`context_annotations.domain`	object	Contains elements which identify detailed information regarding the domain classification based on Tweet text.
`context_annotations.domain.id`	string	Contains the numeric value of the domain.
`context_annotations.domain.name`	string	Domain name based on the Tweet text.
`context_annotations.domain.description`	string	Long form description of domain classification.
`context_annotations.entity`	object	Contains elements which identify detailed information regarding the domain classification bases on Tweet text.
`context_annotations.entity.id`	string	Unique value which correlates to an explicitly mentioned Person, Place, Product or Organization
`context_annotations.entity.name`	string	Name or reference of entity referenced in the Tweet.
`context_annotations.entity.description`	string	Additional information regarding referenced entity.
`entities`	object	Contains details about text that has a special meaning in a Tweet. To return this field, add `tweet.fields=entities` in the request's query parameter.
`entities.annotations`	array	Contains details about annotations relative to the text within a Tweet.
`entities.annotations.start`	integer	The start position (zero-based) of the text used to annotate the Tweet. All start indices are inclusive.
`entities.annotations.end`	integer	The end position (zero based) of the text used to annotate the Tweet. While all other end indices are exclusive, this one is inclusive.
`entities.annotations.probability`	number	The confidence score for the annotation as it correlates to the Tweet text.
`entities.annotations.type`	string	The description of the type of entity identified when the Tweet text was interpreted.
`entities.annotations.normalized_text`	string	The text used to determine the annotation type.
`entities.urls`	array	Contains details about text recognized as a URL.
`entities.urls.start`	integer	The start position (zero-based) of the recognized URL within the Tweet. All start indices are inclusive.
`entities.urls.end`	integer	The end position (zero-based) of the recognized URL within the Tweet. This end index is exclusive.
`entities.urls.url`	string	The URL in the format tweeted by the user.
`entities.urls.expanded_url`	string	The fully resolved URL.
`entities.urls.display_url`	string	The URL as displayed in the Twitter client.
`entities.urls.unwound_url`	string	The full destination URL.
`entities.hashtags`	array	Contains details about text recognized as a Hashtag.
`entities.hashtags.start`	integer	The start position (zero-based) of the recognized Hashtag within the Tweet. All start indices are inclusive.
`entities.hashtags.end`	integer	The end position (zero-based) of the recognized Hashtag within the Tweet.This end index is exclusive.
`entities.hashtags.tag`	string	The text of the Hashtag.
`entities.mentions`	array	Contains details about text recognized as a user mention.
`entities.mentions.start`	integer	The start position (zero-based) of the recognized user mention within the Tweet. All start indices are inclusive.
`entities.mentions.end`	integer	The end position (zero-based) of the recognized user mention within the Tweet. This end index is exclusive.
`entities.mentions.username`	string	The part of text recognized as a user mention. You can obtain the expanded object in `includes.users` by adding `expansions=entities.mentions.username` in the request's query parameter.
`entities.cashtags`	array	Contains details about text recognized as a Cashtag.
`entities.cashtags.start`	integer	The start position (zero-based) of the recognized Cashtag within the Tweet. All start indices are inclusive.
`entities.cashtags.end`	integer	The end position (zero-based) of the recognized Cashtag within the Tweet. This end index is exclusive.
`entities.cashtags.tag`	string	The text of the Cashtag.
`withheld`	object	Contains withholding details for withheld content. To return this field, add `tweet.fields=withheld` in the request's query parameter.
`withheld.copyright`	boolean	Indicates if the content is being withheld for on the basis of copyright infringement.
`withheld.country_codes`	array	Provides a list of countries where this content is not available.
`withheld.scope`	enum (`tweet`, `user`)	Indicates whether the content being withheld is a Tweet or a user.
`public_metrics`	object	Engagement metrics for the Tweet at the time of the request. To return this field, add `tweet.fields=public_metrics` in the request's query parameter.
`public_metrics.retweet_count`	integer	Number of times this Tweet has been Retweeted.
`public_metrics.reply_count`	integer	Number of Replies of this Tweet.
`public_metrics.like_count`	integer	Number of Likes of this Tweet.
`public_metrics.quote_count`	integer	Number of times this Tweet has been Retweeted with a comment (also known as Quote).
`organic_metrics`	object	Organic engagement metrics for the Tweet at the time of the request. Requires user context authentication.
`organic_metrics.retweet_count`	integer	Number of times the Tweet has been Retweeted organically.
`organic_metrics.reply_count`	integer	Number of replies the Tweet has received organically.
`organic_metrics.like_count`	integer	Number of likes the Tweet has received organically.
`promoted_metrics`	object	Engagement metrics for the Tweet at the time of the request in a promoted context. Requires user context authentication.
`promoted_metrics.retweet_count`	integer	Number of times this Tweet has been Retweeted when promoted.
`promoted_metrics.reply_count`	integer	Number of Replies to this Tweet when promoted.
`promoted_metrics.like_count`	integer	Number of Likes of this Tweet when promoted.
`possibly_sensitive`	boolean	Indicates if this Tweet contains URLs marked as sensitive, for example content suitable for mature audiences. To return this field, add `tweet.fields=possibly_sensitive` in the request's query parameter.
`lang`	string	Language of the Tweet, if detected by Twitter. Returned as a BCP47 language tag. To return this field, add `tweet.fields=lang` in the request's query parameter.
`source`	string	The name of the app the user Tweeted from. To return this field, add `tweet.fields=source` in the request's query parameter.
`reply_settings`	string	Shows who can reply to this Tweet. Fields returned are `everyone`, `mentionedUsers`, and `following`. To return this field, add `tweet.fields=reply_settings` in the request's query parameter.
`includes`	object	If you include an `expansion` parameter, the referenced objects will be returned if available.
`includes.tweets`	array	When including the `expansions=referenced_tweets.id` parameter, this includes a list of referenced Retweets, Quoted Tweets, or replies in the form of Tweet objects with their default fields and any additional fields requested using the `tweet.fields` parameter, assuming there is a referenced Tweet present in the returned Tweet(s).
`includes.users`	array	When including the `expansions=author_id` parameter, this includes a list of referenced Tweet authors in the form of user objects with their default fields and any additional fields requested using the `user.fields` parameter.
`includes.places`	array	When including the `expansions=geo.place_id` parameter, this includes a list of referenced places in Tweets in the form of place objects with their default fields and any additional fields requested using the `place.fields` parameter, assuming there is a place present in the returned Tweet(s).
`includes.media`	array	When including the `expansions=attachments.media_keys` parameter, this includes a list of images, videos, and GIFs included in Tweets in the form of media objects with their default fields and any additional fields requested using the `media.fields` parameter, assuming there is a media attachment present in the returned Tweet(s).
`includes.polls`	string	When including the `expansions=attachments.poll_ids` parameter, this includes a list of polls that are attached to Tweets in the form of poll objects with their default fields and any additional fields requested using the `poll.fields` parameter, assuming there is a poll present in the returned Tweet(s).
`errors`	object	Contains details about errors that affected any of the requested Tweets. See Status codes and error messages for more details.
`matching_rules` Default	array	contains the list of filters that matched against the Tweet delivered.
`matching_rules.id` Default	string	ID of the filter rule that matched against the Tweet delivered.
`matching_rules.tag` Default	string	The tag label of the filter rule that matched against the Tweet delivered.

Was this document helpful?

Thank you for the feedback. We’re really glad we could help!

Thank you for the feedback. How could we improve this document?

This page is missing information.

The information was hard to follow or confusing.

There is inaccurate information.

There is a broken link or typo.

Specific Feedback

Thank you for the feedback. Your comments will help us improve our documents in the future.