Xata Has a Free API: A Serverless Database With Built-in Search and AI

#database #typescript #ai #serverless

Xata is a serverless database that combines PostgreSQL, Elasticsearch, and vector search into one platform. It provides a spreadsheet-like UI, full-text search, and AI features without separate services.

Why Xata Matters

Typical stack: PostgreSQL + Elasticsearch + Pinecone + admin panel. Xata replaces all four with one service and one API.

What you get for free:

PostgreSQL-compatible database
Built-in full-text search (Elasticsearch under the hood)
Vector search for AI applications
Ask AI: natural language queries over your data
Spreadsheet-like admin UI
Branching (like git branches for your database)
Free tier: 15GB data, 75 requests/second

Quick Start

npm install -g @xata.io/cli
xata auth login
xata init --db my-app

TypeScript SDK

import { getXataClient } from "./xata";

const xata = getXataClient();

// Create record
const post = await xata.db.posts.create({
  title: "Hello Xata",
  content: "My first post",
  tags: ["intro", "database"],
  published: true,
});

// Query with filters
const posts = await xata.db.posts
  .filter({ published: true })
  .sort("created_at", "desc")
  .getPaginated({ pagination: { size: 20 } });

// Full-text search
const results = await xata.db.posts.search("machine learning", {
  fuzziness: 1,
  prefix: "phrase",
  highlight: { enabled: true },
});

// Vector search (AI)
const similar = await xata.db.posts.vectorSearch("embedding", queryVector, {
  size: 5,
  filter: { published: true },
});

// Ask AI (natural language query)
const answer = await xata.db.posts.ask("What are the most popular topics?", {
  rules: ["Answer based only on the data in the database"],
});
console.log(answer.answer);

Database Branching

# Create branch (like git)
xata branch create staging

# Work on staging branch
xata branch use staging

# Make schema changes, test data
# Then merge back
xata branch merge staging --into main

REST API

# Query records
curl "https://my-workspace.xata.sh/db/my-app:main/tables/posts/query" \
  -H "Authorization: Bearer $XATA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"filter": {"published": true}, "sort": [{"created_at": "desc"}]}'

# Full-text search
curl "https://my-workspace.xata.sh/db/my-app:main/tables/posts/search" \
  -H "Authorization: Bearer $XATA_API_KEY" \
  -d '{"query": "machine learning", "fuzziness": 1}'