AI & Vectors

Supabase provides an open source toolkit for developing AI applications using Postgres and pgvector. Use the Supabase client libraries to store, index, and query your vector embeddings at scale.

The toolkit includes:

• A vector store and embeddings support using Postgres and pgvector.
• A Python client for managing unstructured embeddings.
• An embedding generation process using open source models directly in Edge Functions.
• Database migrations for managing structured embeddings.
• Integrations with all popular AI providers, such as OpenAI, Hugging Face, LangChain, and more.

Search

You can use Supabase to build different types of search features for your app, including:

• Semantic search: search by meaning rather than exact keywords
• Keyword search: search by words or phrases
• Hybrid search: combine semantic search with keyword search

Examples

Check out all of the AI templates and examples in our GitHub repository.

<div className="grid md:grid-cols-12 gap-4 not-prose"> {aiExamples.map((x) => ( <div className="col-span-4" key={x.href}> <Link href={x.href} passHref> <GlassPanel icon={'/docs/img/icons/github-icon'} hasLightIcon={true} title={x.name}> {x.description} </GlassPanel> </Link> </div> ))} </div>

Integrations

<div className="grid md:grid-cols-12 gap-4 not-prose"> {aiIntegrations.map((x) => ( <div className="col-span-4" key={x.href}> <Link href={x.href} passHref> <GlassPanel title={x.name}>{x.description}</GlassPanel> </Link> </div> ))} </div>

Case studies

<div className="grid md:grid-cols-12 gap-4 not-prose"> {[ { name: 'Berri AI Boosts Productivity by Migrating from AWS RDS to Supabase with pgvector', description: 'Learn how Berri AI overcame challenges with self-hosting their vector database on AWS RDS and successfully migrated to Supabase.', href: 'https://supabase.com/customers/berriai', }, { name: 'Firecrawl switches from Pinecone to Supabase for PostgreSQL vector embeddings', description: 'How Firecrawl boosts efficiency and accuracy of chat powered search for documentation using Supabase with pgvector', href: 'https://supabase.com/customers/firecrawl', }, { name: 'Markprompt: GDPR-Compliant AI Chatbots for Docs and Websites', description: "AI-powered chatbot platform, Markprompt, empowers developers to deliver efficient and GDPR-compliant prompt experiences on top of their content, by leveraging Supabase's secure and privacy-focused database and authentication solutions", href: 'https://supabase.com/customers/markprompt', }, ].map((x) => ( <div className="col-span-4" key={x.href}> <Link href={x.href} passHref> <GlassPanel title={x.name}>{x.description}</GlassPanel> </Link> </div> ))} </div>