Meilisearch
App · 12 actions
Search and update Meilisearch indexes: search with filters and sorting, add, read and delete documents, and create and set up indexes. Start a workflow when Meilisearch fails to index something.
Actions
| Action | What it does |
|---|---|
| Add document | Add a document to an index, or change the one with the same ID. |
| Add documents | Add a list of documents to an index, such as rows from an earlier step, changing any with the same ID. |
| Change index settings | Change which fields an index searches, filters and sorts by. |
| Create index | Create an empty index. |
| Delete documents | Delete documents from an index by their IDs, or every document matching a filter. |
| Delete index | Delete an index and every document in it. |
| Get document | Get one document from an index by its ID. |
| Get index | Get an index's document count, fields, and which fields it searches, filters and sorts by. |
| Get task | Get a task a write started: whether it finished, and why it failed if it did. |
| List documents | List an index's documents, optionally only those matching a filter. |
| List indexes | List every index on the server with its ID field. |
| Search | Search an index for documents matching words, with an optional filter and sort. |
Add document
Add a document to an index, or change the one with the same ID.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Document
documentobjectRequired - Its fields and values, including its ID (such as id: 42).
- If it exists
modetext - Whether to change only the fields given or replace the whole document. Defaults to changing the fields given.One of:
update(Add, or update the fields given),replace(Add, or replace the whole document). Default:update. - Wait until done
waittrue/false - Wait up to 30 seconds for Meilisearch to finish, so the next step sees the change. On by default.Default:
true.
Outputs
detailsobject- What it did, such as receivedDocuments and indexedDocuments
durationtext- How long it took (ISO 8601 duration, such as PT0.12S)
enqueued_attext- When it was asked for (ISO 8601)
errortext- Why it failed; empty unless it failed
error_codetext- Meilisearch's code for the failure, such as invalid_document_id
finished_attext- When it finished (ISO 8601); empty while it runs
indextext- The index it works on
statustext- enqueued, processing, succeeded, failed or canceled
task_uidnumber- The task's number, for 'Get task'
typetext- What it does, such as documentAdditionOrUpdate
Example output
{
"details": {
"indexedDocuments": 1,
"receivedDocuments": 1
},
"duration": "PT0.12S",
"enqueued_at": "2026-10-27T10:00:00.000000Z",
"error": "",
"error_code": "",
"finished_at": "2026-10-27T10:00:00.120000Z",
"index": "products",
"status": "succeeded",
"task_uid": 42,
"type": "documentAdditionOrUpdate"
}Add documents
Add a list of documents to an index, such as rows from an earlier step, changing any with the same ID.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Documents
documentslistRequired - The documents, each with its fields and ID. Up to 10000 at a time.
- If one exists
modetext - Whether to change only the fields given or replace the whole document. Defaults to changing the fields given.One of:
update(Add, or update the fields given),replace(Add, or replace the whole document). Default:update. - ID field
primary_keytext - The field that holds each document's ID, for a new index Meilisearch cannot work it out for.
- Wait until done
waittrue/false - Wait up to 30 seconds for Meilisearch to finish, so the next step sees the change. On by default.Default:
true.
Outputs
detailsobject- What it did, such as receivedDocuments and indexedDocuments
durationtext- How long it took (ISO 8601 duration, such as PT0.12S)
enqueued_attext- When it was asked for (ISO 8601)
errortext- Why it failed; empty unless it failed
error_codetext- Meilisearch's code for the failure, such as invalid_document_id
finished_attext- When it finished (ISO 8601); empty while it runs
indextext- The index it works on
statustext- enqueued, processing, succeeded, failed or canceled
task_uidnumber- The task's number, for 'Get task'
typetext- What it does, such as documentAdditionOrUpdate
Example output
{
"details": {
"indexedDocuments": 2,
"receivedDocuments": 2
},
"duration": "PT0.12S",
"enqueued_at": "2026-10-27T10:00:00.000000Z",
"error": "",
"error_code": "",
"finished_at": "2026-10-27T10:00:00.120000Z",
"index": "products",
"status": "succeeded",
"task_uid": 42,
"type": "documentAdditionOrUpdate"
}Change index settings
Change which fields an index searches, filters and sorts by.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Search in
searchable_attributeslist - The fields searched, most important first. Leave empty to keep them.
- Filter by
filterable_attributeslist - The fields a filter may use. Leave empty to keep them.
- Sort by
sortable_attributeslist - The fields results may be sorted by. Leave empty to keep them.
- One result per
distinct_attributetext - A field whose value appears only once in results, such as product_id. Leave empty to keep it.
- Wait until done
waittrue/false - Wait up to 30 seconds for Meilisearch to finish, so the next step sees the change. On by default.Default:
true.
Outputs
detailsobject- What it did, such as receivedDocuments and indexedDocuments
durationtext- How long it took (ISO 8601 duration, such as PT0.12S)
enqueued_attext- When it was asked for (ISO 8601)
errortext- Why it failed; empty unless it failed
error_codetext- Meilisearch's code for the failure, such as invalid_document_id
finished_attext- When it finished (ISO 8601); empty while it runs
indextext- The index it works on
statustext- enqueued, processing, succeeded, failed or canceled
task_uidnumber- The task's number, for 'Get task'
typetext- What it does, such as documentAdditionOrUpdate
Example output
{
"details": {
"filterableAttributes": [
"status"
],
"sortableAttributes": [
"price"
]
},
"duration": "PT0.12S",
"enqueued_at": "2026-10-27T10:00:00.000000Z",
"error": "",
"error_code": "",
"finished_at": "2026-10-27T10:00:00.120000Z",
"index": "products",
"status": "succeeded",
"task_uid": 42,
"type": "settingsUpdate"
}Create index
Create an empty index.
Inputs
- Name
indextextRequired - The new index's name: letters, numbers, '-' and '_', such as products.
- ID field
primary_keytext - The field that holds each document's ID, such as id. Leave empty to let Meilisearch work it out from the first documents.
- Wait until done
waittrue/false - Wait up to 30 seconds for Meilisearch to finish, so the next step sees the change. On by default.Default:
true.
Outputs
detailsobject- What it did, such as receivedDocuments and indexedDocuments
durationtext- How long it took (ISO 8601 duration, such as PT0.12S)
enqueued_attext- When it was asked for (ISO 8601)
errortext- Why it failed; empty unless it failed
error_codetext- Meilisearch's code for the failure, such as invalid_document_id
finished_attext- When it finished (ISO 8601); empty while it runs
indextext- The index it works on
statustext- enqueued, processing, succeeded, failed or canceled
task_uidnumber- The task's number, for 'Get task'
typetext- What it does, such as documentAdditionOrUpdate
Example output
{
"details": {
"primaryKey": "id"
},
"duration": "PT0.12S",
"enqueued_at": "2026-10-27T10:00:00.000000Z",
"error": "",
"error_code": "",
"finished_at": "2026-10-27T10:00:00.120000Z",
"index": "products",
"status": "succeeded",
"task_uid": 42,
"type": "indexCreation"
}Delete documents
Delete documents from an index by their IDs, or every document matching a filter.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Document IDs
document_idslist - The IDs of the documents to delete. Or leave empty and give a filter.
- Filter
filtertext - Instead of IDs, delete every document matching this, such as status = archived.
- Wait until done
waittrue/false - Wait up to 30 seconds for Meilisearch to finish, so the next step sees the change. On by default.Default:
true.
Outputs
detailsobject- What it did, such as receivedDocuments and indexedDocuments
durationtext- How long it took (ISO 8601 duration, such as PT0.12S)
enqueued_attext- When it was asked for (ISO 8601)
errortext- Why it failed; empty unless it failed
error_codetext- Meilisearch's code for the failure, such as invalid_document_id
finished_attext- When it finished (ISO 8601); empty while it runs
indextext- The index it works on
statustext- enqueued, processing, succeeded, failed or canceled
task_uidnumber- The task's number, for 'Get task'
typetext- What it does, such as documentAdditionOrUpdate
Example output
{
"details": {
"deletedDocuments": 1,
"originalFilter": null,
"providedIds": 1
},
"duration": "PT0.12S",
"enqueued_at": "2026-10-27T10:00:00.000000Z",
"error": "",
"error_code": "",
"finished_at": "2026-10-27T10:00:00.120000Z",
"index": "products",
"status": "succeeded",
"task_uid": 42,
"type": "documentDeletion"
}Delete index
Delete an index and every document in it. It cannot be undone.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Wait until done
waittrue/false - Wait up to 30 seconds for Meilisearch to finish, so the next step sees the change. On by default.Default:
true.
Outputs
detailsobject- What it did, such as receivedDocuments and indexedDocuments
durationtext- How long it took (ISO 8601 duration, such as PT0.12S)
enqueued_attext- When it was asked for (ISO 8601)
errortext- Why it failed; empty unless it failed
error_codetext- Meilisearch's code for the failure, such as invalid_document_id
finished_attext- When it finished (ISO 8601); empty while it runs
indextext- The index it works on
statustext- enqueued, processing, succeeded, failed or canceled
task_uidnumber- The task's number, for 'Get task'
typetext- What it does, such as documentAdditionOrUpdate
Example output
{
"details": {
"deletedDocuments": 1
},
"duration": "PT0.12S",
"enqueued_at": "2026-10-27T10:00:00.000000Z",
"error": "",
"error_code": "",
"finished_at": "2026-10-27T10:00:00.120000Z",
"index": "products",
"status": "succeeded",
"task_uid": 42,
"type": "indexDeletion"
}Get document
Get one document from an index by its ID.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Document ID
document_idtextRequired - The document's ID: the value of its ID field.
- Fields to return
fieldslist - Only these fields. Defaults to every field.
Outputs
documentobject- The document's fields
document_idtext- The document's ID
indextext- The index
Example output
{
"document": {
"id": 1,
"price": 79.99,
"status": "published",
"title": "Wireless headphones"
},
"document_id": "1",
"index": "products"
}Get index
Get an index's document count, fields, and which fields it searches, filters and sorts by.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
Outputs
created_attext- When it was made (ISO 8601)
displayed_attributeslist- The fields results show; * for all
distinct_attributetext- The field that keeps results distinct; empty if none
documentsnumber- How many documents it holds
fieldslist- The fields its documents have
filterable_attributeslist- The fields a filter may use
is_indexingtrue/false- True while it is taking in changes
primary_keytext- Its ID field
searchable_attributeslist- The fields it searches; * for all
sortable_attributeslist- The fields results may be sorted by
uidtext- Its name
updated_attext- When it last changed (ISO 8601)
Example output
{
"created_at": "2026-10-01T09:00:00.000000Z",
"displayed_attributes": [
"*"
],
"distinct_attribute": "",
"documents": 1,
"fields": [
"id",
"price",
"status",
"title"
],
"filterable_attributes": [
"status"
],
"is_indexing": false,
"primary_key": "id",
"searchable_attributes": [
"*"
],
"sortable_attributes": [
"price"
],
"uid": "products",
"updated_at": "2026-10-27T10:00:00.120000Z"
}Get task
Get a task a write started: whether it finished, and why it failed if it did.
Inputs
- Task number
task_uidnumberRequired - The task number a write step returned.At least 0.
Outputs
detailsobject- What it did, such as receivedDocuments and indexedDocuments
durationtext- How long it took (ISO 8601 duration, such as PT0.12S)
enqueued_attext- When it was asked for (ISO 8601)
errortext- Why it failed; empty unless it failed
error_codetext- Meilisearch's code for the failure, such as invalid_document_id
finished_attext- When it finished (ISO 8601); empty while it runs
indextext- The index it works on
statustext- enqueued, processing, succeeded, failed or canceled
task_uidnumber- The task's number, for 'Get task'
typetext- What it does, such as documentAdditionOrUpdate
Example output
{
"details": {
"indexedDocuments": 1,
"receivedDocuments": 1
},
"duration": "PT0.12S",
"enqueued_at": "2026-10-27T10:00:00.000000Z",
"error": "",
"error_code": "",
"finished_at": "2026-10-27T10:00:00.120000Z",
"index": "products",
"status": "succeeded",
"task_uid": 42,
"type": "documentAdditionOrUpdate"
}List documents
List an index's documents, optionally only those matching a filter.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Filter
filtertext - Only documents matching this, such as status = draft. The fields must be filterable.
- Fields to return
fieldslist - Only these fields. Defaults to every field.
- How many
limitnumber - How many documents. Defaults to 20.From 1 to 1000.
- Skip
offsetnumber - How many to skip first, to read the next page. Defaults to 0.At least 0.
Outputs
countnumber- How many were returned
documentslist- The documents
indextext- The index
totalnumber- How many match in all
Example output
{
"count": 1,
"documents": [
{
"id": 1,
"price": 79.99,
"status": "published",
"title": "Wireless headphones"
}
],
"index": "products",
"total": 1
}List indexes
List every index on the server with its ID field.
Inputs
None.
Outputs
countnumber- How many indexes
indexeslist- The indexes
Example output
{
"count": 1,
"indexes": [
{
"created_at": "2026-10-01T09:00:00.000000Z",
"primary_key": "id",
"uid": "products",
"updated_at": "2026-10-27T10:00:00.120000Z"
}
]
}Search
Search an index for documents matching words, with an optional filter and sort.
Inputs
- Index
indextextRequired - The index. Pick one, or type its name.
- Search for
querytext - The words to look for. Leave empty to match every document the filter allows.
- Filter
filtertext - Only documents matching this, such as genre = horror. The fields must be filterable.
- Sort by
sortlist - The order of results. The fields must be sortable. Defaults to best match.
- Fields to return
fieldslist - Only these fields of each result. Defaults to every field.
- Match
matchingtext - How many of the words a result must have. Defaults to dropping the last ones if needed.One of:
last(All words, dropping the last ones if needed),all(All words),frequency(All words, dropping the most common ones if needed). Default:last. - How many
limitnumber - How many results. Defaults to 20.From 1 to 1000.
- Skip
offsetnumber - How many to skip first, to read the next page. Defaults to 0.At least 0.
Outputs
countnumber- How many results were returned
estimated_totalnumber- About how many documents match in all
hitslist- The results, best first
indextext- The index searched
processing_msnumber- How long the search took, in milliseconds
querytext- The words searched for
Example output
{
"count": 1,
"estimated_total": 1,
"hits": [
{
"id": 1,
"price": 79.99,
"status": "published",
"title": "Wireless headphones"
}
],
"index": "products",
"processing_ms": 2,
"query": "headphones"
}