Local First WarehouseFree 90 days → paid customer · Closing when full

Turn your Operational Data Store —To an operational context platform for AI agents

Stream, govern, and publish trusted business context that AI agents can safely understand and act on.

Free for 90 daysPick features and your own cloud accountContinuous Support

What's Already Shipping

Operational data is continuously captured from production systems, governed as a data product, and made available to AI agents—without granting direct access to production databases.

pg-cdc - Streams PostgreSQL WAL into typed compacted Parquet

first table - Build semantic models & business entities & features directly from governed operational data

cloud-storage - Keep operational context in low-cost & durable object storage rather than proprietary warehouse platforms

Wire-drop - Securely share governed datasets & operational context & business data products across teams & partners and organizations

Data Lake - Open lake house architecture without warehouse lock-in

Governance - IAM-driven access policies + Sensitivity tagging

Intelligence - AI agents answer business-aware query without production DB access

Versioning model - Git-style analytical workflows for data teams

Iceberg Table - Schema evolution and time travel support

first-table View solutions

Day 0

Apply

5-question form. Takes 3 minutes.

Day 1–3

On Boarding Call

30-min direct call. Confirm fit, walk your stack, align on 90 days.

Day 4–7

White-glove setup

We integrate the platform with your stack. Production-ready by end of week 1

Day 8–90

Build & Ship

Weekly check-ins. Your 2 picked features ship within 90 days.

Eight features. You pick Two. We ship them First.

You're not picking from a wishlist — you're picking which 2 of these planned features we build next.

Foundation

Live

pg-cdc - Postgres Change Data Capture

Moving operational data into S3-compatible object storage

Streams PostgreSQL WAL into typed, compacted Parquet
Creates a physical air gap between production systems and consumers
Pure Go, single binary, no CGO, minimal operational footprint
Immutable CDC history for replay, audit, and recovery
Works with self-hosted PostgreSQL, Amazon Web Services Aurora, and managed cloud databases
Designed for AI-safe access without exposing DATABASE_URL
Pushes directly to S3-compatible object storage

See Details

Analyze

Beta

first-table - Pulls raw Iceberg Table

Curated + State layers → DuckDB → SQL transforms → publishes contracted gold tables

Local-first analytics platform powered by embedded DuckDB
Pull governed Parquet datasets directly to developer laptops
Build analytics, notebooks, semantic models, and features locally
Eliminates always-on warehouse compute costs
Versioned analytical releases for reproducible development
Medallion workflows without distributed infrastructure complexity
Built for teams that want Cloud Warehouse -style workflows without Cloud Warehouse-scale cost

See Details

Storage

Live

Redundant layer -Cloud Storage

Cheap, Centralized, Open Storage

Data lives in low-cost object storage instead of expensive warehouse compute
Open formats: Parquet + Iceberg
Multi-cloud compatible: S3, GCS, Azure Blob, MinIO
Immutable storage architecture improves governance and auditability
Separation of compute and storage by default
Optimized for local-first and edge analytics patterns

See Details

Foundation

Planned

Wire-drop — Secure Data Exchange Subscriber

Secure mTLS data exchange subscriber. External pushes → Iceberg

Secure file and dataset exchange between organizations
Point-to-point governed data delivery
Air-gapped sharing model for regulated environments
Schema-aware ingestion with contract validation
Designed for healthcare, finance, and enterprise B2B data exchange
Lower operational cost than traditional managed transfer systems
Built for “first mile” data ingestion into governed lakes

See Details

Foundation

Live

Data Lake Architecture

Time-travel queries, snapshot isolation, schema evolution. supporting

Raw Layer → immutable CDC + partner data landing zone
Silver Layer → cleaned, normalized, deduplicated datasets
Gold Layer → business-ready analytical models
Exchange Layer → governed external data sharing
Open lakehouse architecture without warehouse lock-in
Designed for AI agents, notebooks, BI tools, and feature pipelines

See Details

Governance

Live

Tag-based ACLs - Untagged data is invisible

Lake Formation sync + full audit trail

Sensitivity tagging at table, column, and dataset level
IAM-driven access policies
Policy enforcement travels with the data
Centralized audit trail for access and transformations
Data contracts instead of undocumented Slack agreements
Built-in support for governed AI access patterns
Lake Formation / catalog integration ready

See Details

Intelligence Layer

Live

YAML contracts - Define schemas, metrics, alerts

YAML contracts define schemas, metrics, alerts.

AI agents read governed datasets, contracts, lineage, and state
Natural language access to operational and analytical data
Business-aware querying without production DB access
Contract-aware agents reduce hallucinations and schema misuse
Context-aware insights across CDC history and gold datasets
Enables secure MCP-compatible AI workflows

See Details

Versioning

Live

CI metrics (authoritative) + analyst observations (positional)

Role-gated annotations. No conflicts.

Git-style analytical workflows for data teams
Versioned datasets and reproducible analytical releases
Controlled schema evolution with audibility
Safe promotion workflows for production analytics
Rollback and replay support using immutable CDC history
PR-style data model evolution workflows

See Details

Why and why now.

Production Postgres data is trapped behind connection strings, replicas, and tribal knowledge.

One-way WAL streaming creates a physical air gap between production systems and downstream AIOpen table formats and cloud object storage eliminate vendor lock-in while enabling scalable analytics with DuckDB, Iceberg, and modern data toolingEvery developer and AI workflow gets isolated, local-first analytical environments powered by governed production data snapshots.Governance travels with the data — IAM policies, lineage, and access controls are enforced without direct database connectivity.

How partnership works

From application to production in under a week. Then 90 days of building together.

Start your application See burnside-project on GitHub Connect on LinkedIn

5-question form. 3-minute commitment

After 90 days: you pivot to a paid plan

You move to managed service platform — with the layer you picked already live for your team. No surprise pricing. No lock-in. You can export your Iceberg data and walk away at any time.

Free for 90 days

Direct founder Slack/email access

White-glove onboarding and setup

One-command Iceberg export workflows

Early access to Starter and Enterprise platform features

Continued founder support for 90 days after onboarding

Explore current repo surfaces Schedule a meeting

Free for 90 days

Choose one platform layer — PostgreSQL CDC, governed datasets, local notebooks, Iceberg workflows, or AI-safe analytics — and we’ll help integrate it into your stack

White-glove onboarding

white-glove onboarding direct founder Slack/email access fast feedback loops roadmap influence early access to new platform features

Get Started →

What We Ask

a quick weekly 30-minute feedback session production or staging PostgreSQL usage optional permission to list your logo on /partners

See →

Who This Is For

dev-first SaaS teams analytics and data engineering teams indie hackers with production apps engineering teams dealing with hot-table queries teams exploring safe AI access to operational data

See a Demo →

Partners signed

We're brand new. You'd be partner #1. (Update this number when partners sign.)

See a Demo →

White-Glove Onboarding

Hands-on help integrating PostgreSQL, partner datasets, and governed local analytics into your stack

AI-ready datasets Book a Meeting