Keyword and vector search in one index, answered in single-digit milliseconds.
No clusters to run or tune. An idle index costs only its storage.
- Hybrid search in one index. BM25 keyword and kNN vector search in the same request, merged with reciprocal rank fusion.
- Fully serverless. No nodes, shards or replicas to size. No upgrades, patching or rebalancing. Scales to zero, and traffic spikes are handled automatically.
- Drop-in for Elasticsearch, OpenSearch, Solr and turbopuffer. Keep your clients, SDKs and queries, any version, and change the endpoint. A query that can't be translated exactly fails with a clear error instead of returning wrong results.
- Built for agents. The WarpSearch MCP server lets coding agents sign you up, create collections, load documents and run searches. You approve one email.
- Your cloud or ours. Runs on AWS, Google Cloud and Azure in multiple regions, or inside your own account with Bring Your Own Cloud.
- Pay for what you use. Billing is based on data stored, writes and searches, with volume discounts applied automatically. Every account gets 1 GB in one collection free, permanently.
With your coding agent. Add the MCP server, then ask your agent to set up search for your data:
claude mcp add --transport http warpsearch https://api.warpsearch.io/mcp # Claude Code
codex mcp add warpsearch --url https://api.warpsearch.io/mcp # Codex
gemini mcp add --transport http warpsearch https://api.warpsearch.io/mcp # Gemini CLIOver the API. Create an API key in the console, then run a hybrid query:
curl -X POST https://api.warpsearch.io/v1/collections/products/search \
-H "Authorization: Bearer $WARPSEARCH_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"language": "es",
"query": {
"query": { "match": { "title": "coffee" } },
"knn": { "field": "embedding", "query_vector": [0.85, 0.15, 0.05, 0.05], "k": 10 }
},
"limit": 10
}'From Elasticsearch, OpenSearch, Solr or turbopuffer. Point your existing client at WarpSearch:
| You use | Endpoint |
|---|---|
| Elasticsearch | https://api.warpsearch.io/compat/es |
| OpenSearch | https://api.warpsearch.io/compat/opensearch |
| Solr | https://api.warpsearch.io/compat/solr |
| turbopuffer SDKs (Python, TypeScript, Go, Java, Ruby) | https://api.warpsearch.io/compat/turbopuffer |
Moving from turbopuffer needs no code change when your settings come from the environment:
export TURBOPUFFER_BASE_URL=https://api.warpsearch.io/compat/turbopuffer
export TURBOPUFFER_API_KEY=ws_live_...
unset TURBOPUFFER_REGIONThe migration guide and turbopuffer compatibility pages cover what works unchanged and how to cut over. The WarpSearch API also accepts SQL, PostgreSQL (with pgvector distances), Lucene and LogQL. See Query languages.
Looking for an alternative? Side-by-side pricing, latency and features:
| Instead of | Compare |
|---|---|
| turbopuffer | WarpSearch vs turbopuffer: flat price per request instead of per GB queried, and Elasticsearch, OpenSearch and Solr APIs on the same data |
| Elasticsearch | Elasticsearch Serverless · Managed Elasticsearch |
| OpenSearch | OpenSearch Serverless · Managed OpenSearch · Aiven for OpenSearch |
| Vector databases | Pinecone · Weaviate · Amazon S3 Vectors |
| Hosted search | Algolia |
| Repository | What it is |
|---|---|
| WarpSearch.Umbraco.Search | WarpSearch provider for Umbraco Search: full-text search, filters and facets for Umbraco sites without a search cluster or Examine indexes. On NuGet. |
Quickstart · Vector search · Hybrid search · MCP for agents · API reference · Pricing · Compare
Questions or a larger deployment in mind? Contact us or write to hello@warpsearch.io.