Skip to content
Integration Runtime Layer

search

Full-text + vector search — pgvector (Postgres). BM25 keyword search and vector cosine search over your own database, no external search infra.

packages/integrations/search · Source ↗ · Docs ↗

  • Billing v2 migration playbook

    0.94

    Step-by-step migration from legacy invoices to the new metering pipeline.

    posts/2024/billing-v2.md
    x

  • Migration checklist for billing tenants

    0.89

    Verify RLS, secrets rotation, and webhook backfill before cutover.

    docs/billing/migration.mdx
    x

  • Creem + WeChat Pay billing adapter notes

    0.81

    One billing contract keeps customer code stable across both rails.

    changelog/2026/adapters.md
    x

  • Zero-downtime migration patterns

    0.76

    Dual-write, shadow-read, then flip — applies to billing data too.

    docs/patterns/migration.mdx
    x
Searched 12,401 docs in 38ms · Postgres backend38ms

In code

search.ts
1import { getSearch } from "@nebutra/search";
2
3// Postgres pgvector + full-text — no separate search cluster to run.
4const search = await getSearch();
5
6await search.indexDocument("posts", {
7  id: post.id,
8  title: post.title,
9  body: post.body,
10  tenantId: post.tenantId,
11});
12
13const results = await search.search("posts", {
14  query: "billing migration",
15  tenantId: org.id,
16});

More in nebutra.com

All 17 packages