Conversation threads using the X API
If you look at how conversations evolve on X, one Post can spark several conversation threads, each of which can grow in length and complexity as more people chime in. Identifying relationships between Posts and understanding conversation threads is a feature of the X API v2 payload and search capabilities. When Posts are posted in response to a Post (known as a reply), or in response to a reply, there is now a defined conversation_id on each reply, which matches the Post ID of the original Post that started the conversation.
Replies to a given Post, as well as replies to those replies, are all included in the conversation stemming from the single original Post. Regardless of how many reply threads result, they will all share a common conversation_id to the original Post that sparked the conversation. Using the X API v2, you have the ability to retrieve and reconstruct an entire conversation thread, so that you can better understand what is being said, and how conversations and ideas evolve.
The example below shows a conversation thread of five different people, including one reply to a reply:
{
"data": [{
"conversation_id": "1279944223114900000",
"in_reply_to_user_id": "1102323333",
"author_id": "63044444",
"created_at": "2020-07-06T15:58:10.000Z",
"id": "1280169177479744444",
"referenced_tweets": [{
"type": "replied_to",
"id": "1280155225706433333"
}],
"text": "@ThirdPerson333 @OriginalPerson000 Reply to the third reply!"
},
{
"conversation_id": "1279944223114900000",
"in_reply_to_user_id": "3001960000",
"author_id": "1102323333",
"created_at": "2020-07-06T15:02:44.000Z",
"id": "1280155225706433333",
"referenced_tweets": [{
"type": "replied_to",
"id": "1279944223114900000"
}],
"text": "@OriginalPerson000 Third reply"
},
{
"conversation_id": "1279944223114900000",
"in_reply_to_user_id": "3001960000",
"author_id": "199562222",
"created_at": "2020-07-06T15:02:36.000Z",
"id": "1280155190306340864",
"referenced_tweets": [{
"type": "replied_to",
"id": "1279944223114900000"
}],
"text": "@OriginalPerson000 Second Reply"
},
{
"conversation_id": "1279944223114900000",
"in_reply_to_user_id": "3001960000",
"author_id": "179201111",
"created_at": "2020-07-06T01:10:15.000Z",
"id": "1279945722494811111",
"referenced_tweets": [{
"type": "replied_to",
"id": "1279944223114900000"
}],
"text": "@OriginalPerson000 First Reply"
}
],
"includes": {
"users": [{
"name": "Original person",
"id": "3001960000",
"username": "OriginalPerson000"
},
{
"name": "First person",
"id": "179201111",
"username": "FirstPerson111"
}, {
"name": "Second person",
"id": "199562222",
"username": "SecondPerson222"
}, {
"name": "Third person",
"id": "1102323333",
"username": "ThirdPerson333"
}, {
"name": "Fourth person",
"id": "63044444",
"username": "FourthPerson444"
}
],
"tweets": [{
"conversation_id": "1279944223114900000",
"in_reply_to_user_id": "3001960000",
"author_id": "1102323333",
"created_at": "2020-07-06T15:02:44.000Z",
"id": "1280155225706433333",
"referenced_tweets": [{
"type": "replied_to",
"id": "1279944223114900000"
}],
"text": "@OriginalPerson000 Third reply"
},
{
"conversation_id": "1279944223114900000",
"author_id": "3001960000",
"created_at": "2020-07-06T01:04:17.000Z",
"id": "1279944223114900000",
"text": "This is the original post"
}
]
},
"meta": {
"newest_id": "1280169177479744444",
"oldest_id": "1279945722494811111",
"result_count": 4
}
}
Retrieving conversation_id as a tweet.fields parameter
To request the conversation_id for all Posts returned on a v2 endpoint, the tweet.fields=conversation_id field can be added to the request parameters. The conversation_id field is always the Post ID of the original Post in the conversation reply thread. All Posts within the same reply thread, including reply threads that are created from earlier reply threads, will show the same conversation_id.
Request with conversation_id parameter
curl --request GET \
--url 'https://api.twitter.com/2/tweets?ids=1225917697675886593&tweet.fields=author_id,conversation_id,created_at,in_reply_to_user_id,referenced_tweets&expansions=author_id,in_reply_to_user_id,referenced_tweets.id&user.fields=name,username' \
--header 'Authorization: Bearer $ACCESS_TOKEN'
Response
{
"data": [
{
"id": "1225917697675886593",
"text": "@TwitterEng *ahem* https://t.co/aroJHt2zQ1",
"created_at": "2020-02-07T23:02:10.000Z",
"author_id": "2244994945",
"in_reply_to_user_id": "6844292",
"conversation_id": "1225912275971657728",
"referenced_tweets": [
{
"type": "quoted",
"id": "1200517737669378053"
},
{
"type": "replied_to",
"id": "1225912275971657728"
}
]
}
],
"includes": {
"users": [
{
"username": "TwitterDev",
"name": "Twitter Dev",
"id": "2244994945"
},
{
"username": "TwitterEng",
"name": "Twitter Engineering",
"id": "6844292"
}
],
"tweets": [
{
"id": "1200517737669378053",
"text": "|￣￣￣￣￣￣￣￣￣￣￣|\n don't push \n to prod on \n Fridays \n|＿＿＿＿＿＿＿＿＿＿＿| \n(\\__/) ||\n(•ㅅ•) ||\n/ づ",
"created_at": "2019-11-29T20:51:47.000Z",
"author_id": "2244994945",
"conversation_id": "1200517737669378053"
},
{
"id": "1225912275971657728",
"text": "Note to self: Don't deploy on Fridays",
"created_at": "2020-02-07T22:40:37.000Z",
"author_id": "6844292",
"conversation_id": "1225912275971657728"
}
]
}
}
Using conversation_id as a filter operator
The conversation_id can be used as a search query parameter when using either recent search or as an operator within a rule for filtered stream. Using the operator on its own will result in the entire conversation thread of Posts being returned in either real time through filtered stream, or paginated in reverse chronological order from search Tweets. You can also receive a count of the Posts in a conversation using this operator with Posts counts.
Additional operators can be added to the query/rule, in conjunction with the conversation_id operator, however these will apply only to the Posts that are part of that conversation. Only one conversation_id can be specified at a time without an OR clause within the query/rule.
Reconstructing the conversation can be done by ordering the Posts with a matching conversation_id by timestamp, and taking note of which Posts are directly in reply to other Posts in the conversation thread. This can be accomplished by also requesting the in_reply_to_user_id field and referenced_tweets.id and in_reply_to_user_id expansions.
Request to query by conversation_id
curl --request GET \
--url 'https://api.twitter.com/2/tweets/search/recent?query=conversation_id:1279940000004973111&tweet.fields=in_reply_to_user_id,author_id,created_at,conversation_id' \
--header 'Authorization: Bearer $ACCESS_TOKEN'
Response
Note: Results for search Posts are in reverse chronological order.
{
"data": [{
"id": "1280169000000704333",
"text": "@attributeisland @iterationjoe What beautiful creatures! Happy #seaturtleweek",
"conversation_id": "1279940000004973111",
"public_metrics": {
"retweet_count": 0,
"reply_count": 0,
"like_count": 7,
"quote_count": 0
}
},
{
"id": "1280166000000519222",
"text": "@attributeisland \"One touch of nature makes the whole world kin.\" -John Muir",
"conversation_id": "1279940000004973111",
"public_metrics": {
"retweet_count": 0,
"reply_count": 1,
"like_count": 1,
"quote_count": 0
}
},
{
"id": "1280166000000519221",
"text": "@attributeisland I love turtles!",
"conversation_id": "1279940000004973111",
"public_metrics": {
"retweet_count": 1,
"reply_count": 0,
"like_count": 4,
"quote_count": 0
}
},
{
"id": "1280166000000519220",
"text": "@attributeisland Turtlemoji🐢",
"conversation_id": "1279940000004973111",
"public_metrics": {
"retweet_count": 0,
"reply_count": 0,
"like_count": 1,
"quote_count": 0
}
},
{
"id": "1279940000004973111",
"text": "Sea turtles are roaming in our waters!",
"conversation_id": "1279940000004973111",
"public_metrics": {
"retweet_count": 67,
"reply_count": 11,
"like_count": 396,
"quote_count": 2
}
}
],
"meta": {
"newest_id": "1280169000000704333",
"oldest_id": "1279940000004973111",
"result_count": 5
}
}