What is a unified API?
A unified API provides a single, consistent interface to interact with multiple third-party services. Instead of learning each provider’s unique schema, you define one schema that works across all of them. This guide covers one of Ampersand’s use cases: building unified APIs. You can also use Ampersand for provider-specific integrations or mix both approaches. Unlike traditional unified API providers that force you into their predefined models, Ampersand lets you design your own schema.Why build a unified API?
- Faster development - Add new integrations without changing application logic
- Consistent data model - One schema regardless of source
- No vendor lock-in - You control the schema and where data lives
How it works
1. Map objects and fields
2. Receive unified + raw data
Every webhook includes your mapped fields, non-mapped fields, and the complete provider response:3. Real-time updates with subscribe actions
Get webhooks when data changes (1-2 minutes latency):Key architectural differences from Merge
Data persistence
Ampersand: No data storage for better security posture. Ampersand acts as the transport and transformation layer only. You control where and how to persist data. Merge: Stores all customer data in Merge servers by default. “Merge Destinations” (Enterprise) enables you to turn off storage. Why it matters: For healthcare, finance, or regulated industries, Ampersand’s transport-only architecture eliminates compliance concerns around third-party data storage.Two-way sync
Ampersand:- Read: Scheduled syncs or real-time subscribe actions with field-level granularity
- Write: Direct writes with automatic field mapping - your unified schema works for both reads and writes
- Pattern: Subscribe + write = true bidirectional real-time sync
- Read: Scheduled syncs + poll after webhook notifications (model-level)
- Write: Direct writes, but no unified mapping - you must query
/metaper integration to discover required fields - Pattern: Polling + writes
Comparison table
| Feature | Ampersand | Merge |
|---|---|---|
| Schema | You define it | Fixed common model |
| Raw data | Always included | Requires include_remote_data=true flag |
| Data storage | No persistent storage | Stored in Merge cloud |
| Real-time | Subscribe actions (all plans) | Third-party webhooks (Pro/Enterprise) |
| Event granularity | Field-level changes | Model-level notifications |
| Write mapping | Automatic - reuses your read schema | Manual - query /meta per integration |
Ticketing unified API
Standardize tickets across Jira, Linear, Zendesk, and ServiceNow:Migration from Merge
Understand the persistence model
Merge: Stores your customer data; you poll their API
Ampersand: Streams data to your webhooks; you store itSet up webhook endpoints and decide on your storage strategy (database, warehouse, etc.)
Ampersand: Streams data to your webhooks; you store itSet up webhook endpoints and decide on your storage strategy (database, warehouse, etc.)
Complete examples
Related resources
- Object and Field Mapping - Technical details on mapping configuration
- Read Actions - Scheduled data syncs and backfills
- Subscribe Actions - Real-time event subscriptions
- Write Actions - Writing data back to providers
- Destinations - Configuring webhook destinations

