FireIgniteFlow Documentation

The definitive guide to real-time NoSQL to Relational data streaming. Learn how our Go ETL Engine infers schemas and handles complex Firebase topologies.

1. Architecture Overview

FireIgniteFlow operates completely transparently. We listen to your Firestore Write-Ahead Logs (via Firebase Admin Streams) and map them directly to your Postgres instance in under 50ms.

Client AppWrites JSON Data

Writes

Firestore NoSQLWrite-Ahead Logs

Stream

ETL Go Engine< 50ms Processing

Pipeline

PostgreSQLNormalized tables

2. Complex Schema Inference

Handling nested JSON maps and arrays is where most ETLs fail. FireIgniteFlow automatically flattens objects and creates normalized relational join tables for arrays.

Firestore Document (NoSQL)

// Collection: users
{
  "id": "usr_99x",
  "metadata": {
    "lastLogin": "2026-05-23T10:00:00Z",
    "device": "iOS"
  },
  "roles": ["admin", "editor"]
}

Auto-Generated Postgres DDL

-- Main Table (Flattened Maps)
CREATE TABLE users (
  id VARCHAR PRIMARY KEY,
  metadata_lastLogin TIMESTAMPTZ,
  metadata_device VARCHAR
);

-- Auto-Normalized Array Table
CREATE TABLE users_roles (
  user_id VARCHAR REFERENCES users(id),
  value VARCHAR
);

3. Handling Deletions (Tombstones)

When a document is deleted in Firebase, FireIgniteFlow does NOT delete the row in Postgres. Instead, we use the Tombstone Pattern to preserve historical integrity.

A system column _fs_deleted_at is automatically added to all your Postgres tables.

SELECT * FROM users
WHERE _fs_deleted_at IS NULL;

Ready to supercharge your Analytics?

FireIgniteFlow is currently onboarding high-volume startups. Apply for early access to get your Firebase data flowing into Postgres in milliseconds.

Apply for Closed Beta