db9 — Postgres but for agents

db9

          // Create a database with vector support
          $ db9 db create --name rag-app
          Database created: abc123
           
          // Upload docs to cloud filesystem (for RAG)
          $ db9 fs cp ./docs/ abc123:/knowledge/
          Uploaded 42 files to /knowledge/
           
          // Create embeddings table with pgvector
          $ db9 db sql abc123 -q "CREATE TABLE docs (
            id serial PRIMARY KEY,
            content text,
            embedding vector(1536),
            metadata jsonb
          )"
          CREATE TABLE
           
          // Add HNSW index for fast similarity search
          $ db9 db sql abc123 -q "CREATE INDEX ON docs USING hnsw (embedding vector_cosine_ops)"
          CREATE INDEX
           
          // Semantic search with cosine similarity
          $ db9 db sql abc123 -q "SELECT content, 1 - (embedding <=> $1) AS score
            FROM docs ORDER BY embedding <=> $1 LIMIT 5"
          content                              score
          ─────────────────────────────────────  ─────
          Authentication uses JWT tokens...    0.94
          API rate limits are per-customer...  0.87
          Webhooks require HTTPS endpoints...  0.82
          (3 rows)
        

Features

Everything you need

From database creation to production monitoring, db9 covers the full lifecycle.

⚡

Instant databases

Spin up a serverless Postgres instance in seconds. No provisioning, no config files.

🤖

Vector search

pgvector built-in with HNSW indexes. Semantic search, RAG, and embeddings — no external vector DB needed.

📑

SQL from CLI

Execute queries inline, from files, or pipe from stdin. Output as table, JSON, or CSV.

🌱

Database branching

Create schema branches for testing and development. Isolated environments in one command.

📈

Built-in observability

Inspect QPS, latency, slow queries, and connection metrics without external tools.

🛠

Migration management

Create, apply, and track SQL migrations. Integrated status tracking per database.

📝

Type generation

Generate TypeScript or Python types from your database schema automatically.

🔍

JSONB + GIN indexes

Store and query semi-structured data with lightning-fast GIN indexes. Perfect for agent memory and tool outputs.

📦

Dump & seed

Export schemas and data as SQL. Seed databases from files for reproducible environments.

📁

Cloud filesystem

Read, write, and manage files per database via the SDK or interactive shell. Built for RAG pipelines and file-driven workflows.

CLI Reference

30+ commands

A comprehensive toolkit for every database operation.

db9 db create --name myapp

Create a serverless Postgres database

db9 db sql <id> -q "..."

Execute SQL queries directly

db9 db branch create <id> --name dev

Branch a database for dev/test

db9 db inspect <id>

View QPS, latency, and metrics

db9 gen types <id> --lang typescript

Generate TS/Python types from schema

db9 migration up <id>

Apply pending migrations

db9 db dump <id> -o backup.sql

Export database as SQL

db9 db connect <id>

Get connection string for psql

db9 fs sh <id>

Interactive filesystem shell

db9 fs cp local.txt <id>:/data/

Upload files to cloud filesystem

Postgresbut for agents