In FalconeMulti-tenant Backend-as-a-Service

Postgres & MongoDB data APIs, object storage, events, serverless functions and realtime — behind one gateway, isolated per tenant, deployable to Kubernetes, OpenShift or air-gapped clusters.

What is In Falcone?

Quickstart (TODO app)

Architecture

🏢

Tenant isolation by construction

Every data path is scoped by tenant. PostgreSQL Row-Level Security + a non-BYPASSRLS application role, MongoDB adapter-injected tenant filters, and per-tenant realtime channels. Cross-tenant access fails closed.

Security model

🗄️

Postgres & MongoDB data APIs

REST data access over both engines with keyset pagination, filtering and DDL. One control plane, two storage shapes.

Data API

🔑

Anon & service API keys

Supabase-style flc_anon_… and flc_service_… keys routed at the gateway, plus per-tenant JWT issuance. Embed read-only access in a frontend safely.

Gateway & keys

⚡

Realtime subscriptions

Server-Sent Events backed by MongoDB change streams and PostgreSQL trigger-based CDC, tenant-scoped inside the pipeline.

Realtime

📦

Storage, events & functions

S3-compatible object storage (MinIO), an event bus (Kafka/Redpanda) and serverless functions round out the BaaS surface.

Components

🚀

Deploy anywhere

A single umbrella Helm chart with layered values for Kubernetes (Ingress), OpenShift (Routes, restricted-v2) and air-gapped private registries — plus a docker-compose stack for local development.

Installation

🤖

Built for AI — a BaAIS

A backend designed to be natively consumable by AI agents — MCP server hosting (Preview, served live under /v1/mcp) and the Temporal-based Flows workflow engine (Preview) — all under the same per-tenant isolation, auth and quotas.

MCP & Flows

Not production-ready

In Falcone is in early, active development. Public APIs, data schemas and runtime behavior may change at any time, without notice or a migration path; there are no stability, security or support guarantees; and the project has not undergone a security audit. Do not use it for production workloads or sensitive data — evaluation, experimentation and development only. See the Roadmap.

A first look

In Falcone ships with a web console for tenant administration and data exploration. The screenshots below are from a real deployment running on a Kubernetes cluster.

Tenants overview