Bring Vector Search
to Vercel AI SDK

Connect your Redis database (equipped with the RediSearch module) to your AI agent, turning it into an advanced Vector Database administrator. Activating this integration grants your conversational interface the power to interact directly with your semantic search engine, enabling tasks like querying mathematical embeddings for similar records, configuring fresh vector indexes, and managing geometric data structures without needing dedicated external database clients.

What you can do

• Similarity Vector Search (KNN) — Let the AI perform rapid native vector comparisons (search_vectors). Provide an embedding array via prompt or code, and retrieve the absolute nearest top_k neighbors securely cached in your infrastructure.
• Index Management — Actively discover all loaded RediSearch vector indexes, investigate their configured dimensions (get_index_info), or command the AI to instantiate new KNN indexes (create_vector_index) tailored for fresh AI workloads.
• Embedding Administration — Inject and modify geometric vector components associated with a document key (upsert_vector), or purge legacy embeddings efficiently (delete_vector) to keep semantic records clean and operational.

How it works

1. Authorize the Redis Vector MCP connector from your module catalog.
2. Configure it securely by providing your full Redis URL (ensure it points to a Redis instance that natively supports RediSearch vector extensions).
3. Prompt your AI to "find the top 5 nearest neighbors for this JSON array in the 'products-index'" or "create a new 1536-dimensional vector index for OpenAI embeddings."

Who is this for?

• AI & ML Engineers — Rapidly iterate over similarity tuning. Store resulting chunk embeddings on the fly, and query KNN vectors right from the prompt instead of scripting Python drivers repeatedly.
• Backend Developers — Maintain semantic storage logic. Audit schemas, map out active index properties, and delete obsolete hashes holding raw vector models instantly.
• Data Architects — Validate your Redis vector environments interactively. Explore dimension structures and index readiness confirming architecture viability for RAG (Retrieval-Augmented Generation) applications.