- Search conversations
You can search for multiple conversations by the value of their attributes in order to fetch exactly which ones you want.
To search for conversations, you need to send a POST request to https://api.intercom.io/conversations/search.
This will accept a query object in the body which will define your filters in order to search for conversations.
Search queries can be complex, so optimizing them can help the performance of your search. Use the AND and OR operators to combine multiple filters to get the exact results you need and utilize pagination to limit the number of results returned. The default is 20 results per page and maximum is 150. See the pagination section for more details on how to use the starting_after param.
You can nest these filters in order to get even more granular insights that pinpoint exactly what you need. Example: (1 OR 2) AND (3 OR 4). There are some limitations to the amount of multiple's there can be:
- There's a limit of max 2 nested filters
- There's a limit of max 15 filters for each AND or OR group
Most keys listed in the conversation model are searchable, whether writeable or not. The value you search for has to match the accepted type, otherwise the query will fail (ie. as created_at accepts a date, the value cannot be a string such as "foorbar"). The source.body field is unique as the search will not be performed against the entire value, but instead against every element of the value separately. For example, when searching for a conversation with a "I need support" body - the query should contain a = operator with the value "support" for such conversation to be returned. A query with a = operator and a "need support" value will not yield a result.
| Field | Type |
|---|---|
| id | Integer |
| type | String |
| created_at | Date (Unix timestamp in seconds) |
| updated_at | Date (Unix timestamp in seconds) |
| source.type | String Accepted fields are conversation, email, facebook, instagram, phone_call, phone_switch, push, sms, twitter and whatsapp. |
| source.id | Integer |
| source.delivered_as | String |
| source.subject | String |
| source.body | String |
| source.author.id | String |
| source.author.type | String |
| source.author.name | String |
| source.author.email | String |
| source.url | String |
| contact_ids | String |
| teammate_ids | String |
| admin_assignee_id | Integer |
| team_assignee_id | Integer |
| channel_initiated | String |
| open | Boolean |
| read | Boolean |
| state | String |
| waiting_since | Date (Unix timestamp in seconds) |
| snoozed_until | Date (Unix timestamp in seconds) |
| tag_ids | String |
| priority | String |
| statistics.time_to_assignment | Integer |
| statistics.time_to_admin_reply | Integer |
| statistics.time_to_first_close | Integer |
| statistics.time_to_last_close | Integer |
| statistics.median_time_to_reply | Integer |
| statistics.first_contact_reply_at | Date (Unix timestamp in seconds) |
| statistics.first_assignment_at | Date (Unix timestamp in seconds) |
| statistics.first_admin_reply_at | Date (Unix timestamp in seconds) |
| statistics.first_close_at | Date (Unix timestamp in seconds) |
| statistics.last_assignment_at | Date (Unix timestamp in seconds) |
| statistics.last_assignment_admin_reply_at | Date (Unix timestamp in seconds) |
| statistics.last_contact_reply_at | Date (Unix timestamp in seconds) |
| statistics.last_admin_reply_at | Date (Unix timestamp in seconds) |
| statistics.last_close_at | Date (Unix timestamp in seconds) |
| statistics.last_closed_by_id | Integer |
| statistics.count_reopens | Integer |
| statistics.count_assignments | Integer |
| statistics.count_conversation_parts | Integer |
| conversation_rating.requested_at | Date (Unix timestamp in seconds) |
| conversation_rating.replied_at | Date (Unix timestamp in seconds) |
| conversation_rating.score | Integer |
| conversation_rating.remark | String |
| conversation_rating.contact_id | String |
| conversation_rating.admin_id | String |
| ai_agent_participated | Boolean |
| ai_agent.resolution_state | String |
| ai_agent.last_answer_type | String |
| ai_agent.rating | Integer |
| ai_agent.rating_remark | String |
| ai_agent.source_type | String |
| ai_agent.source_title | String |
The table below shows the operators you can use to define how you want to search for the value. The operator should be put in as a string ("="). The operator has to be compatible with the field's type (eg. you cannot search with > for a given string value as it's only compatible for integer's and dates).
| Operator | Valid Types | Description |
|---|---|---|
| = | All | Equals |
| != | All | Doesn't Equal |
| IN | All | In Shortcut for OR queries Values most be in Array |
| NIN | All | Not In Shortcut for OR ! queries Values must be in Array |
| > | Integer Date (Unix timestamp in seconds) | Greater than |
| >= | Integer Date (Unix timestamp in seconds) | Greater than or equal to |
| < | Integer Date (Unix timestamp in seconds) | Lower than |
| <= | Integer Date (Unix timestamp in seconds) | Lower than or equal to |
| ~ | String | Contains |
| !~ | String | Doesn't Contain |
| ^ | String | Starts With |
| $ | String | Ends With |
Tickets are only included in the results if the access token has the read_tickets scope. Without it, tickets are left out and no error is returned.
Results are sorted by updated_at, newest first. To sort differently, add a sort object to the request body with a field and an order of ascending or descending. field is required whenever sort is sent, and must be one of the fields in the table above, other than type, state, source.subject, source.body, source.author.id, source.author.type, source.author.name, source.author.email, conversation_rating.remark, ai_agent.rating_remark and ai_agent.source_title, which can't be sorted. A missing field, or one that isn't searchable, returns a 400 error with the code invalid_field.
Intercom API version.
If you omit the header, the API uses the version set in your app package, and 1.0 if none is set.
Search using Intercoms Search APIs with a single filter.
- The production API serverhttps://api.intercom.io/conversations/search
- The european API serverhttps://api.eu.intercom.io/conversations/search
- The australian API serverhttps://api.au.intercom.io/conversations/search
curl -i -X POST \
'https://api.intercom.io/conversations/search?include_monitors=true&include_scorecards=true' \
-H 'Authorization: Bearer <YOUR_TOKEN_HERE>' \
-H 'Content-Type: application/json' \
-H 'Intercom-Version: Preview' \
-d '{
"query": {
"operator": "AND",
"value": [
{
"field": "created_at",
"operator": ">",
"value": "1306054154"
}
]
},
"pagination": {
"per_page": 5
}
}'successful
The list of conversation objects
Cursor-based pagination is a technique used in the Intercom API to navigate through large amounts of data. A "cursor" or pointer is used to keep track of the current position in the result set, allowing the API to return the data in small chunks or "pages" as needed.
- successful
- Response with `include_monitors=true` and `include_scorecards=true`
{ "type": "conversation.list", "pages": { "type": "pages", "page": 1, "per_page": 5, "total_pages": 1 }, "total_count": 1, "conversations": [ { … } ] }