The local database that should have existed.

Store JSON, query it, analyze it — all in-engine. No daemon, no schema, no ETL. Just npm install.

Rust core, mmap-backed storage, ~2µs reads.

$ npm install flintdb

demo.ts

import  FlintDB  from "flintdb";

const db = FlintDB.open("./mydata");

db.put("stocks", 
  ticker: "AAPL",
  close: 195.5,
  volume: 5400000,
);

const result = db.query(
  from: "stocks",
  filter:  op: "eq", field: "ticker", value: "AAPL" ,
  sort:  field: "close", descending: true ,
);

What you don't need anymore.

No more schema headaches

JSON in, JSON out. API responses, logs, nested objects — store them as-is. No migrations, no schema definitions, no ORM.

No more `await` for local data

Your data is right here on disk. FlintDB’s native API is synchronous and completes in ~2 microseconds. No async overhead for local operations.

No more ‘CSV it first’ for analysis

avg, stddev, EMA, percentile, correlation — computed inside the engine. Run analytics where your data lives. No Pandas, no ETL pipeline.

No more separate vector DB

Store embeddings alongside your documents. Cosine and L2 search built in. One database for data and vectors.

Query is Data

Queries are plain JSON objects — no SQL strings, no query parser.

db.query(
  from: "prices",
  filter: 
    op: "and",
    filters: [
       op: "eq", field: "ticker", value: "AAPL" ,
       op: "gte", field: "date", value: "2025-01-01" ,
    ],
  ,
  sort:  field: "date", descending: true ,
  limit: 100,
);

db.query(
  from: "prices",
  group_by: ["ticker"],
  select: [
     name: "ticker", type: "col", field: "ticker" ,
     name: "avg_close", type: "fn", func: "avg", field: "close" ,
     name: "volatility", type: "fn", func: "stddev", field: "close" ,
     name: "max_volume", type: "fn", func: "max", field: "volume" ,
  ],
  sort:  field: "avg_close", descending: true ,
);

db.query(
  from: "prices",
  filter:  op: "eq", field: "ticker", value: "AAPL" ,
  window:  order_by: "date", rows: 20 ,
  select: [
     name: "date", type: "col", field: "date" ,
     name: "close", type: "col", field: "close" ,
     name: "sma_20", type: "fn", func: "avg", field: "close" ,
     name: "ema_20", type: "fn", func: "ema", field: "close" ,
     name: "vol", type: "fn", func: "stddev", field: "close" ,
  ],
);

How FlintDB compares

Feature	FlintDB	SQLite	DuckDB	MongoDB
Setup	npm install	Native addon or binary	npm install (WASM)	Separate server
Query Language	JSON objects	SQL strings	SQL strings	JSON (server required)
Built-in Analytics	stddev, EMA, percentile, correlation...	Manual SQL	Rich SQL analytics	Aggregation pipeline
Vector Search	Built-in (cosine, L2)	Extension needed	Extension needed	Atlas Search (cloud)
Architecture	In-process, sync	In-process	In-process	Client-server

Simple pricing

Core

Free

Collections, Indexes, Views, Aliases
Filter, Sort, GroupBy, Window
avg, sum, stddev, percentile, ema, correlation
Vector Search (Cosine, L2)

Get Started

Pro

$99/yr

per module

Everything in Core, plus:

Financial Indicators (MACD, RSI, Bollinger, Ichimoku, VWAP)
Statistical Analysis (t-test, ANOVA, Chi-square, K-means)
Seasonal Decompose, Anomaly Detection
Linear Forecast

Get Pro

Get started in seconds

npm $ npm install flintdb

pip $ pip install flintdb

GitHub Documentation Quick Start