Platform Overview

A unified operational data engine for live operational state.

Tapstate Platform is a unified operational data engine that captures, transforms, and serves live operational state.

Start with reliable log-based CDC. Expand into transformation and state serving in the same product path.

Production systems

Committed source changes

Tapstate Platform

ConnectorsOrderingCheckpointsRecovery
Capture
Transform
Serve

Operational State Layer

current · queryable · reusable

ApplicationsAPIsAutomationAI agents

HOW TEAMS BUILD A DATA PATH

Define, validate, and run one declarative data path.

Tapstate uses .tapstate.yml resources to define connections, apply supported transforms, and declare delivery targets. Teams can review the configuration in Git, validate it before a run, and test connectivity and data behavior in their own environment.

Step 1

Define connections

Choose connector roles, capture modes, endpoints, and credentials.

Step 2

Define a pipeline

Reference a source, apply supported transforms, and declare sync or push targets.

Step 3

Review and validate

Check resource structure, fields, IDs, references, and secret placeholders.

Step 4

Run in your environment

Test connectivity and representative data behavior before production.

Capture → Transform → Serve

One engine. Three responsibilities.

Tapstate keeps capture, ordered transformation, and state delivery on the same operational path.

Tapstate Platform

Capture

Load existing data and, where supported, continue from source changes.

SnapshotFull load + CDCSource positions + checkpointsConnector-specific recovery

Transform

Apply ordered, reviewable transforms to records in flight.

Filter recordsJavaScript enrichmentRename + normalizeArray unwind + fan-out

Serve

Materialize and deliver current state through supported targets.

Structured syncUpsert or appendDeletes by stable keySchema-change policy

Shared across the path

One deployableDeclarative resourcesWorkspace validationDurable recovery state

Progressive adoption

Start with Capture. Expand on the same product path.

Teams can begin with capture and delivery. As requirements grow, the same product path can add ordered transformation and materialize current state through supported targets.

Capture

Load existing data or continue from supported source changes.

Capture + Transform

Add ordered transforms to filter, normalize, enrich, and reshape records.

Capture + Transform + Serve

Materialize consumer-ready current state through supported targets.

Start with the responsibilities the use case needs today.

Expand on the same product path as requirements grow.

The output

Maintain current state once. Deliver it to many consumers.

Tapstate materializes source changes into consumer-ready current state through supported targets. Applications, APIs, automation, and AI agents can reuse the same maintained result without rebuilding the full data path.

Source-oriented changes

  • order row updated
  • payment inserted
  • shipment status changed
  • customer tier updated
  • support case opened

Tapstate Platform

Capture → Transform → Serve

Current order state

  • order status
  • payment state
  • shipment position
  • customer tier
  • active support context

Downstream consumers

Applications

Use current entities.

APIs

Expose maintained state.

Automation

Act on current conditions.

AI agents

Use coherent context.

What Tapstate operates

The full operating path, not just one component.

Tapstate maintains the definitions, progress, policies, and diagnostic context required to operate a data path across capture, transformation, and delivery.

  1. 01

    Declarative resources

    Keep source, pipeline, transform, and delivery definitions explicit and reviewable.

  2. 02

    Positions and checkpoints

    Retain capture progress, checkpoints, and retry context across interruptions.

  3. 03

    Transform and schema state

    Apply ordered rules while keeping mappings and output shape inspectable.

  4. 04

    Delivery policies

    Control target writes, stable keys, deletes, and schema-change behavior.

  5. 05

    Validation and diagnostics

    Validate resources and retain status, retry, and failure context for troubleshooting.

Product boundary

One engine with a clear operating boundary.

Tapstate operates the data path from source capture through transformation and delivery. It works alongside the systems that remain authoritative, analytical, or responsible for business behavior.

Inside Tapstate

Capture modes and progress

Connector-defined reads, source positions, and checkpoints.

Ordered transformation

Reviewable transform rules, mappings, and output shape.

Materialization and delivery

Target writes, stable keys, deletes, and schema-change policies.

Validation and recovery context

Resource validation, diagnostics, retries, and durable operating state.

Tapstate works alongside

Systems of record

Remain authoritative for operational facts and source-side behavior.

Event backbones

Continue to support custom event distribution and specialized stream topologies.

Warehouses and lakehouses

Retain historical data for analysis, aggregation, and model training.

Applications and agent frameworks

Own business logic, user experience, decisions, and execution.

Explore the Architecture

Quickstart

Build and validate a real data path.

The Tapstate quickstart defines a MySQL source, a MongoDB target connection, and a pipeline that links them. Validate the workspace before testing connectivity and data behavior in a non-production environment.

Source · MySQL · Snapshot

orders_source.tapstate.yml

Defines the MySQL source connection and snapshot read mode.

Pipeline

orders_to_customer_view.tapstate.yml

References the source and declares the MongoDB delivery target.

Target · MongoDB

customer_view.tapstate.yml

Defines the MongoDB target connection.

Validate the workspace

$tapstate validate --workdir tapstate-work
Open the Quickstart

Next steps

See how Tapstate fits your operational data path.

Review the architecture or walk through your use case with us.