Event Streaming & Event-Driven Architecture at apis.io/use-cases/event-streaming/ resolves 50 providers and 1,101 APIs — event backbones and message brokers that carry events between producers and consumers, plus the decoupled architectures built on top: webhooks, choreography, event sourcing.
It is the largest use-case by API count on the network.
The cohort
| Provider | Score | Band | APIs |
|---|---|---|---|
| Salesforce | 80.5 | Exemplar | 294 |
| Convoy | 73.6 | Exemplar | 11 |
| Oracle GoldenGate | 71.3 | Exemplar | 37 |
| Amazon Kinesis | 70.8 | Exemplar | 11 |
| Google Forms | 69.8 | Strong | 3 |
| Azure Container Apps | 69.1 | Strong | 5 |
| Amazon Managed Apache Flink | 68.1 | Strong | 31 |
| Adobe Experience Cloud | 67.9 | Strong | 36 |
| Azure Functions | 67.9 | Strong | 23 |
| Choreo | 67.1 | Strong | 16 |
| Amazon MSK | 65.7 | Strong | 6 |
| AWS Lambda | 65.6 | Strong | 12 |
| Amazon MQ | 65.5 | Strong | 5 |
A build path
The cohort splits into three layers, and choosing at the wrong layer is the usual mistake.
Transport. Amazon MSK (65.7) for managed Kafka when you need a durable, replayable log and your team already thinks in topics. Amazon Kinesis (70.8) when you want the same shape with less operational surface. Amazon MQ (65.5) when you are carrying an existing JMS or AMQP application and need a broker rather than a log.
Processing. Amazon Managed Apache Flink (68.1) for genuine stream processing — windowing, joins across streams, stateful aggregation. AWS Lambda (65.6) and Azure Functions (67.9) for the far more common case where each event is handled independently and you just need code to run.
Delivery. Convoy (73.6) is the specialist worth knowing about — 11 APIs, exemplar band, and it does one thing: reliable webhook delivery with retries, signing, and a dead-letter path. Most teams build that badly in-house three times before buying it.
Change data capture. Oracle GoldenGate (71.3) when the events you need are database mutations rather than application-emitted messages.
The Salesforce entry
Salesforce at 80.5 across 294 APIs is the highest composite in this cohort and slightly misleading as a “streaming” provider. It appears here because of Platform Events and Change Data Capture — a CRM emitting a genuine event stream. If your events originate in Salesforce, that stream is the source of truth and you should consume it directly rather than polling.
The description gap, again
This week’s AsyncAPI piece found 1,768 definitions across just 661 providers on a catalog of more than 25,000. That gap is felt precisely here. The 50 providers in this use-case are the ones who build event infrastructure — they describe events well because events are their product.
Everyone else sends webhooks and documents them in prose. The infrastructure is excellent; the description of what flows through it is where the catalog thins out.
Takeaway
50 providers, 1,101 APIs, and a clean three-layer split between transport, processing, and delivery. Pick the layer before the vendor — and if all you actually need is reliable webhook delivery, Convoy at 73.6 will save you from building it yourself.
Browse the use-case at apis.io/use-cases/event-streaming/.