Searching Streams - Upstash Documentation

Upstash Redis Search can index the entries of a Redis stream. Each stream entry becomes a separate search document, so you can search event fields without scanning the stream or knowing an entry ID.

How stream indexes work

A stream index differs from a JSON, hash, or string index in two ways:

It is bound to one exact stream key instead of one or more key prefixes.
Each stream entry is a document. The document ID returned by Search is the stream entry ID.

The fields added with XADD provide the document fields. Stream payloads are flat field-value pairs, so schema fields refer to stream fields rather than nested paths.

The stream does not need to exist when you create the index. Entries added later are indexed automatically.

Create a stream index

Use ON STREAM followed by the exact stream key. PREFIX is not supported for stream indexes.

SEARCH.CREATE event-search ON STREAM events SCHEMA message TEXT service KEYWORD severity U64 FAST occurredAt DATE FAST

Only one search index can be bound to a stream at a time. To bind the stream to a different index, first drop the existing index with SEARCH.DROP.

Add and search entries

Add entries with regular stream commands. The field names do not need to appear in the schema, but only schema fields are searchable.

XADD events 1-0 message "Payment authorization failed" service checkout severity 4 occurredAt 2026-08-24T09:30:00Z
XADD events 2-0 message "Order completed" service checkout severity 1 occurredAt 2026-08-24T09:31:00Z

SEARCH.WAITINDEXING event-search

SEARCH.QUERY event-search '{"message":"authorization","severity":{"$gte":3}}' SELECT 3 message service severity

The matching result has 1-0 as its document ID. Without NOCONTENT, its content is returned as the stream entry’s field-value pairs. All regular query features, including filtering, highlighting, sorting, pagination, and score functions, use the fields declared in the schema.

Index updates are asynchronous. Use SEARCH.WAITINDEXING when a query must include preceding writes, especially in tests and setup scripts.

Schema values

Redis stores stream field values as strings. Search converts each value according to its schema type:

U64, I64, and F64 fields require values in the corresponding numeric format.
BOOL fields accept true or false.
DATE fields require an RFC 3339 timestamp, such as 2026-08-24T09:30:00Z.
TEXT, KEYWORD, and FACET fields use string values.

Missing fields and values that cannot be converted are omitted from the indexed document. If an entry has no valid schema fields, it is not added to the index. You can use FROM to expose a stream field under a different index field name:

SEARCH.CREATE event-search ON STREAM events SCHEMA description TEXT FROM message severity U64 FAST

Queries use the schema name (description), while SELECT and returned content use the original stream field name (message).

Existing entries and reindexing

Creating an index scans entries already present in the stream. Use SKIPINITIALSCAN to index only new entries at first:

SEARCH.CREATE event-search SKIPINITIALSCAN ON STREAM events SCHEMA message TEXT severity U64 FAST

Run SEARCH.REINDEX later to rebuild the index from all entries currently in the stream.

Entry deletion and trimming

The index follows the contents of the stream:

Stream change	Search index change
`XADD`	Adds a document for the new entry. Entries removed by an `XADD` trim are removed too.
`XDEL`, `XDELEX`, or `XACKDEL`	Removes the documents for deleted entries.
`XTRIM`	Removes the documents for trimmed entries.
`DEL`, `UNLINK`, or expiration of the stream key	Removes all documents but keeps the index definition.

If a deleted or expired stream is recreated under the same key, new entries are added to the existing index. Consumer-group operations that do not change entry data, such as XACK, do not change the index.

RENAME, RENAMENX, and COPY are not supported for stream keys.

Next steps

See Schema Definition for field types and options.
See Querying and filtering for the JSON query syntax.
See Aggregations to group and summarize stream entries.

​How stream indexes work

​Create a stream index

​Add and search entries

​Schema values

​Existing entries and reindexing

​Entry deletion and trimming

​Next steps

How stream indexes work

Create a stream index

Add and search entries

Schema values

Existing entries and reindexing

Entry deletion and trimming

Next steps