Throughline
One connector interface, every system in sync.
A multi-connector ETL platform that syncs six vendor systems into PostgreSQL and Airtable behind one shared connector contract.
- language
- Python 3.11, Next.js dashboard
- connectors
- 6 vendor integrations sharing one contract
- interface
- 4-method connector ABC + plugin registry
- ownership
- Sole engineer — design, build, and production operation
- in production
- A PR + podcast company; actively maintained
- history
- built Oct 2025 to Jul 2026; in production since
- repo
- Private production repo; public extract: throughline-connector-kit
The problem
A PR and podcast company ran on four disconnected SaaS tools with no unified view of a client. Throughline put the systems behind one shared connector contract feeding a single source of truth.
Stakeholder signal
Finance and ops staff at a PR and podcast business — client-facing teams answer billing and delivery questions faster with one contact layer instead of contradictory records across systems.
How it's built
-
One interface, every integration for free
Every connector implements the same four methods and registers through a decorator. A new source inherits incremental sync, retry queues, health checks, and dual-destination writes without touching the sync engine. The shared contract keeps source-specific logic at the connector boundary while the engine owns incremental sync, retries, health checks, and dual-destination writes.
-
Incremental sync with per-connector checkpoints
A sync-metadata table records the last successful timestamp per connector and entity type, so each run fetches only what changed. The architecture docs report a 69% reduction in daily API operations; the public basis is the documented comparison of 3,395 incremental calls against an 11,548-call full-refresh baseline.
-
A Postgres MDM layer for contact identity
A master_contacts table uses generated columns to normalize email, phone, and name, and cross-links QuickBooks, Copper, Basecamp, and PandaDoc identifiers into a single row, so a contact is one entity instead of four disconnected records.
By the numbers
self-reported marks figures stated in docs or commit history that the source brief couldn't reproduce from the repository alone. Everything else traces back to code.
- 69%: Basis: Throughline architecture docs compare 3,395 daily API operations with an 11,548-operation full-refresh baseline; no raw production export is published.
Where this honestly stands today
Shipped to production. Six vendor integrations share the connector contract; sync semantics, checkpoints, retries, and health checks remain in the engine. Efficiency gains are documented from production operations rather than exported as raw telemetry.
The production repo stays private. The sanitized public extract shows the four-method connector protocol, in-memory sync engine, and tests — no credentials, vendor schemas, or client data.
Want the parts that aren't in a public repo? Happy to walk you through the architecture and the decisions on a call.