Burnside Project

Open infrastructure for turning operational data into AI-ready business context.

Most AI agents can read documents.

Very few can understand how your business actually works.

Burnside helps organizations capture operational activity from production systems, govern it as trusted business context, and safely expose it to AI agents.

What Can You Build?

Customer Support Agents

Customer: "Where is my order?"

Instead of searching multiple systems, an AI agent can access governed order history, shipment status, inventory, and support interactions.

Powered by:

pg-cdc
Business Context Models
ai-dial-pad

Customer Success Copilots

"What customers are at risk of churn?"

An AI agent can combine subscription activity, support tickets, usage trends, and account history to help customer success teams prioritize outreach.

Powered by:

pg-cdc
first-table
ai-dial-pad

Sales Assistants

"Which opportunities are most likely to close this quarter?"

AI agents can reason over operational data, customer activity, product adoption, and account history without direct production database access.

Powered by:

pg-cdc
first-table

Operations Copilots

"What changed yesterday?"

AI agents can monitor operational activity, inventory movement, order delays, and fulfillment bottlenecks across the business.

Powered by:

pg-cdc
ai-dial-pad

Executive Intelligence Agents

"Give me a morning briefing."

AI agents summarize business performance using governed operational context rather than manually assembled reports.

Powered by:

Business Context Models
Operational Context Publishing

Projects

pg-cdc

Capture operational activity from PostgreSQL and publish governed, immutable data products.

first-table

Transform operational data into business entities, semantic models, and AI-ready context.

wire-drop

Securely exchange governed business context across teams, partners, and organizations.

ai-dial-pad

Publish business context as AI-ready endpoints for ChatGPT, Claude, and enterprise agents.

Architecture

Production Systems → Capture → Govern → Model → Publish → AI Agents

No direct AI access to production databases.

Just trusted business context.

Build Locally. Scale Intentionally.

Our Projects - For turning operational data into AI-ready business context.

pg-cdc

A single-binary PostgreSQL change-data-capture daemon that streams WAL into typed Parquet files on cloud storage. Pure Go, no CGO, no Kafka, no JVM. Produces an immutable, governed data zone that downstream consumers (Athena, Spark, Iceberg-aware engines, AI agents via the MCP server) read from without ever touching the source database

https://github.com/burnside-project/pg-cdc

pg-warehouse

The data platform is built from 5 independent projects across 4 layers. Each layer has a clear responsibility and communicates via Iceberg tables in a shared Glue catalog.

https://github.com/burnside-project/pg-warehouse

wire-drop (in private beta)

Secure data exchange subscriber. Publishers push data over mTLS; wire-drop receives, deduplicates, writes Iceberg tables, and enforces tag-based governance.

ai-dial-pad(in Private Beta)

Publish governed data products as dial-able tiny-URL MCP endpoints — paste the URL into Claude or ChatGPT Enterprise and start a governed, audited conversation with your data.

pg-collector

pg-collector is a lightweight edge compute agent that extracts PostgreSQL telemetry, processes it locally through a DuckDB analytical warehouse, and delivers Parquet files to our cloud platform where AI analyzes patterns and predicts issues before they impact your users.

Single binary. Zero runtime dependencies. YAML config. Runs anywhere — systemd, Docker, Kubernetes, bare metal.

https://github.com/burnside-project/pg-collector

Why We Exist

Modern analytics stacks have become expensive, complex, and difficult to govern.

Organizations spend hundreds of thousands of dollars per year on data warehouses, ETL pipelines, governance tools, and AI integrations.