cat open-source.md

Answer any agent question on Parquet.

Scalable search + analytics data infrastructure on object storage, for 10x cheaper.

Contact an engineer Get started for free

ClickBench: Infino ranks #1 among search engines for SQL analytics →

trusted by

United States Government
Pendo
Bazinga Labs

Built by the creators of OpenSearch

cat query.sql

Any agent question can be answered on Infino.

1An agent asks a question

How manyA count. The agent wants a number. SQL does this with COUNT. errors last nightFilters. Keep the error-level rows and the last 24 hours. mention disk fullKeyword search. Match those exact words in the log body. Today this is a call to a search engine. or look like a capacity incidentSemantic search. Match meaning even when the log never says disk full. Today this is a call to a vector database., and are tied to last week’s outageGraph search. Keep the errors linked to the outage’s record through a key they share, like a host or a service, even when they never mention it. Today this is a call to a graph database.? Which serviceA group on the same hits. SQL GROUP BY. Today a second warehouse query after the searches return. had them, and which team owns itA join. Match each service to its owner in a second table. Today the agent looks this up in another system and stitches the answers together.?

2It hides seven jobs

count
group
join
filter
keyword search
semantic search
graph search

hybrid search →

3Infino SQL runs all seven

query.sql

SELECT   o.team, h.service, count(*) AS hits
FROM     hybrid_search(
           'logs', 'body', 'disk full',
           'embedding', :q, 1000
         ) AS h
JOIN     graph_walk('logs', 'incidents', [4521], 2, 5000) AS g
           ON g._id = h._id
JOIN     owners o ON o.service = h.service
WHERE    h.level = 'error'
  AND    h.ts > now() - interval '24 hours'
GROUP BY o.team, h.service
ORDER BY hits DESC;

infino migrate

Replace five systems with Infino and run on one copy of your data.

01
Start with search
Swap out Lucene under OpenSearch, or load an Elasticsearch snapshot. Your clients and Dashboards stay as they are.
02
Add the next workload
Vectors, graph and analytics land in the same Parquet files in your bucket.
03
Consolidate everything
Retire Pinecone, Neo4j, ClickHouse and Snowflake one at a time, until one copy serves every agent.

migrate from Elasticsearch or OpenSearch →

infino bench

Faster, cheaper and more scalable than any other retrieval engine in the world.

Keyword · 1B documents

Infino top-10 latency: 78 ms
Lucene top-10 latency: 536 ms

Vector · 1M vectors

Infino p99 latency at recall ≥ 0.95: 1.1 ms
OceanBase p99 latency at recall ≥ 0.95: 1.6 ms
Elasticsearch p99 latency at recall ≥ 0.95: 1.9 ms
Milvus p99 latency at recall ≥ 0.95: 2.0 ms
Redis p99 latency at recall ≥ 0.95: 2.4 ms

SQL · ClickBench

Infino SQL score, lower is better: 1.55×
OpenSearch SQL score, lower is better: 16.40×
Quickwit SQL score, lower is better: 19.39×
Elasticsearch SQL score, lower is better: 28.52×
SigLens SQL score, lower is better: 90.11×

Analytics cost · 1 PB

Infino monthly bill: $401K
Databricks monthly bill: $1.30M
Snowflake monthly bill: $1.85M

Search cost · 1 PB

Infino monthly bill: $398K
OpenSearch monthly bill: $5.74M
Elasticsearch monthly bill: $6.34M

benchmarks →

infino diff

How Infino compares.

Capability	Infino	Databricks	Snowflake	Elastic
Storage	Parquet in your bucket	Delta, plus a vector index	Snowflake or Iceberg, plus a search index	Its own index
Search	In the same files	Vector Search endpoints	Cortex Search service	Lucene
SQL and joins	Yes	Yes	Yes	Limited (ES\|QL)
Context for agents	Table cards and join keys	Genie spaces	Semantic models	Agent Builder
Cost at 1 PB	1×	3.2×	4.6×	15.9×
Answer accuracy	73.3%	60.0%	57.8%	n/a

infino spec

Specs in the cloud or on-prem.

architecture →

search: full-text · vector · hybrid · graph · NL
index: BM25 (PFOR-delta, FST) · HNSW · OPANN + Sq16
engine: Rust + Small Language Models (SLMs)
language: SQL (Apache DataFusion) · REST · Query DSL · MCP · NL
storage: object storage, S3 · GCS · Azure Blob · on-prem
format: Apache Parquet
deploy: on-prem · self-hosted cloud · hosted cloud
core license: Apache-2.0
security: SOC 2 Type II
encryption: TLS in transit · AES-256 at rest

infino init

Queries, made simple.