LangSearch
Web search and semantic reranking APIs that add relevant external context to LLM applications.
About LangSearch
LangSearch provides APIs for web retrieval and semantic reranking in LLM applications. The Web Search API returns context from web documents, including news, images, and videos, while hybrid keyword and vector retrieval supports natural-language queries. The Semantic Rerank API accepts an existing candidate set, measures the relationship between each document and the query, and returns results ordered by relevance. It can serve as either a primary source of web context or a second-stage relevance layer for BM25, vector, or hybrid search pipelines. Individuals and small teams can use the service free of charge, with no fees or subscriptions stated. Integration still requires engineering work, and public detail is limited for teams evaluating security, privacy, support, procurement, or larger-scale deployment.
Best for
Best use cases
Use the Web Search API when an LLM workflow needs context from current web documents, news, images, or videos.
Send an initial result set to the Semantic Rerank API when first-stage retrieval needs relevance refinement.
Combine keyword and vector retrieval when exact matching and semantic intent both matter to the search experience.
Compare before choosing
Compared with a standalone keyword search layer, LangSearch adds vector retrieval and semantic reranking options.
Compared with building a reranker in-house, the API supplies document relevance scores through a hosted endpoint.
The free offer lowers initial cost, but public plan detail is limited for larger organizations evaluating scale or support needs.
Features
Ground LLM responses
Retrieve current web context for an LLM application through a dedicated search API.
Rerank retrieval results
Reorder keyword, vector, BM25, or RRF candidates using semantic relevance scores.
Build hybrid search
Combine keyword and vector retrieval for applications that need both exact and semantic matching.
Pros
- Free access removes upfront API fees for individuals and small teams testing retrieval workflows.
- Web search and semantic reranking can cover both initial retrieval and second-stage relevance refinement.
- Compatibility with keyword, vector, BM25, and RRF results makes the reranker usable in existing search stacks.
Cons
- API-first delivery requires engineering work for authentication, integration, evaluation, and production monitoring.
- The public pricing offer gives limited decision detail for larger teams with scale, support, or procurement requirements.
- No dedicated security or privacy overview is presented in the product coverage, limiting trust evaluation for sensitive deployments.
Pricing Plans
Free access
- Web Search API access
- Semantic Rerank API access
- Access intended for individuals and small teams
* Pricing information is for reference only. Please refer to the latest announcement on LangSearch's official website.