Pusher vs Stream

Discover the differences between Pusher and Stream, their pros and cons, and which is best to use for your chat or messaging solution.

If you're building realtime chat functionality into your application, there are many tools that promise to get you to market faster. Two in particular, Pusher and Stream (previously GetStream), take quite different approaches to helping you build realtime chat functionality.

That can make it hard to choose the right solution for your app and for your team. Do you want an out-of-the-box chat platform that offers less flexibility or do you want more control at the price of having to write and maintain more of your own code?

In this article, we'll aim to make that decision easier by comparing how Pusher and Stream differ in terms of what they do, the guarantees they offer, and how they charge for their services.

Copy link to clipboard

Pusher vs Stream high level overview

The details differ quite considerably between Pusher and Stream and that stems from one core reason. Both Pusher and Stream set out to solve different problems:

  • Pusher is a general purpose messaging platform that helps you distribute messages using WebSocket.

  • Stream is an API and platform that specializes in higher level functionality including chat, video and audio, and activity feeds.

The impact is that Stream can help you solve a limited range of problems but with less custom development on your part, whereas Pusher is open ended but requires more effort. 

So, which should you choose? First, we need to look in more detail at what both products offer.

Copy link to clipboard

What is Pusher?

Pusher, which is part of the MessageBird suite of tools, is a general purpose messaging PaaS and offers two distinct products:

Pusher Beams is pretty straightforward. It gives you a single API to deliver push notifications to iOS, Android, and web users. Pusher Channels is a little more involved, offering a platform for two-way messaging between clients and a central Pusher backend.

There's some room for confusion here because "message" has a different meaning in the context of chat versus realtime systems, so let's clear that up before we move on. For Pusher, messages are self-contained packets of data that publishers, such as one of your backend services or the client on an end-user's mobile device, emits to a topic, which is then consumed by subscribers in the form of other end-users and services. 

Pusher's job is to take care of consuming and delivering messages to the right place. But does it do a good job? We can get a quick overview by reviewing Pusher's pros and cons.

Copy link to clipboard

Advantages of Pusher

  • Use case flexibility: Pusher is a general purpose pub/sub messaging platform. That means you can use it to create a wide range of event-driven applications.

  • Available in many languages: Pusher provides SDKs for frontend clients, including web and mobile platforms, as well as several backend frameworks, including Rails, Node.js, and .NET.

  • Protocol fallback: Pusher's protocol uses WebSocket, meaning it's available just about everywhere. However, if there's a problem with WebSocket, Pusher can also switch to HTTP streaming and HTTP polling.

  • Compliance and security: Pusher can encrypt private channels and it complies with GDPR and ISO27001. While Pusher doesn't comply with HIPAA by default, the company does offer guidance on how to achieve HIPAA compliance using their platform.

Copy link to clipboard

Disadvantages of Pusher

  • Pusher provides low-level tooling: Building a chat solution on top of Pusher will take more work than using a dedicated chat platform as a service (PaaS) tool like Stream. That means more code to write and maintain.

  • Weak guarantees: A PaaS like Pusher has one job: to get your messages to their destination. Unfortunately, Pusher doesn't guarantee that your messages will reach their destination. Instead, it says that they will be delivered "under normal system operation". That means you need to build your own message tracking functionality. There's also no guarantee over message ordering, which could lead to messages arriving in the wrong order.

  • No message history: There are times when clients disconnect or otherwise can't receive their intended messages. In those cases, you'll need some other mechanism to store message history because Pusher doesn't store messages.

  • Limited to one region: Pusher routes all of your data through a single location, meaning that it is vulnerable to outages and other degradations in service. That also impacts user experience because end-users further away from the cloud region will necessarily experience higher latencies.

Copy link to clipboard

What is Stream?

Stream takes a different approach to Pusher. Rather than giving you the building blocks to create your own solutions, Stream offers use-case specific APIs and infrastructure for:

  • Chat

  • Audio and video calling

  • Activity feeds

  • Automatic moderation

  • AI chat bot integration

Rather than providing lower-level realtime infrastructure, Stream aims to give you an out-of-the-box tool to solve specific problems. That distinguishes it from Pusher in two key ways: it does much more for you but offers less flexibility.

We should get a better understanding of what makes Stream a good option and where it has weaknesses.

Copy link to clipboard

Advantages of Stream

  • Time to market: Stream's website promises to help you build a chat app in days. If Stream's solutions meet your needs, it's likely that you can get your chat functionality to market far quicker than with a generalized tool like Pusher.

  • Little ongoing maintenance: Because Stream provides chat functionality out of the box, it means there's less original code and infrastructure for you to maintain than when using Pusher. 

  • Edge network: By running a global network of servers, your end-users could experience lower latencies with Stream when compared to the single cloud location offered by Pusher. But that's dependent on the geographic location of your users and other factors, such as the efficiency of the code running within Stream's network.

  • High capacity: Stream claims to support up to 5 million concurrent connections in a single channel.

  • Strong uptime if you pay for it: With a promise of 99.999% uptime for enterprise customers, Stream promises a reliable solution if you are willing to pay higher prices. Stream offers a relatively unambitious 99.95% uptime for lower-priced packages, which is equivalent to more than four hours of downtime each year.

  • Complies with standards: Stream Chat and Feeds comply with SOC 2 and HIPAA. 

  • Ready-made chat UI kits and other integrations: Stream gives you a headstart with chat UI components and backend SDKs for various frameworks and languages.

Copy link to clipboard

Disadvantages of Stream

  • Limited to specific use cases: Stream isn't a general purpose realtime platform. That might make it easier to get set up initially if you need chat or activity feeds - but as soon as your needs no longer align with what Stream has to offer then you'll either have to migrate to a more flexible platform or change your product plans.

  • Can be expensive: The convenience of Stream's ready made solutions comes at a price. The cheapest paid-for plan is $499 per month (or $399 per month if paid annually), supporting 10,000 monthly active users (MAU). However, things get really expensive once you exceed that user limit with a charge of 9 cents per month for each additional MAU and 99 cents for each additional concurrently connected user. If you were to meet Stream's claimed 5 million concurrent user capacity then you would pay an overage fee of $5,398,605 per month, based on list prices.

  • Lack of data ownership: When working with Stream, your chat data must reside in their database. That could cause issues depending on how sensitive the data is and whether it is covered by regulations that Stream does not meet. It could also be harder to integrate with third-party systems, such as observability platforms or data stores, as you have less access to the data underlying Stream's services.

Copy link to clipboard

Pusher vs Stream feature comparison

As we've seen, Pusher and Stream can both help you to build chat functionality into your application. However, the approach they each take differs greatly. Of course, the most obvious difference is that Stream provides you with a tailored chat solution, while Pusher gives you greater flexibility to build any kind of messaging based functionality. But there are other differences, too, in terms of how they perform and the developer experience they offer.

Let's look at them both, side-by-side.

Copy link to clipboard

Performance, reliability, scalability, and messaging guarantees

Both platforms are only useful if they perform when you need them. So, can you rely on Pusher and Stream?

Feature/capability

Pusher

Stream

Uptime guarantee

Pusher no longer publishes an uptime guarantee but, previously, it offered a 99.95% SLA (equivalent to 21 minutes downtime every month).

99.95% SLA for lower tiers.

99.999% for enterprise plans.

Latency

Pusher doesn't publish latency stats but their materials mention latencies of under 240 ms.

Stream reports a median latency of 170ms to connect an Amsterdam-based user with someone in the US East and a 99th percentile of 359ms.

Scalability

Pusher's pricing suggests that enterprise plans can scale to 10 million or more concurrent connections.

Pusher claims to scale to billions of messages.

Stream reports having connected 5 million concurrent users to a single chat channel in testing. However, it's not clear what numbers Stream can support in production.

Infrastructure

9 cloud regions.

Global edge network with cloud regions in five locations.

Multi-region

No. You must choose one region.

No. Despite its edge  network, which handles CORS validation and other handshake-type requests, chat functionality runs in one region.

Guaranteed message ordering

If you send each message individually, they will arrive in a random order. However, you can batch messages together, in which case they will be delivered at the same time and in order, albeit with greater average latency than if each message were sent individually.

The order of messages within a group or channel is guaranteed. If messages are being delivered to a webhook, their order is also preserved.

Guaranteed message delivery (exactly-once)

No

Unknown

Automatic reconnections with continuity

Pusher supports automatic reconnections  but you'll need to store and load message history within your own code.

Unknown.

Stream's developer documentation doesn't address reconnections.

Copy link to clipboard

Chat features

Chat users expect rich features, such as typing indicators, reactions, emojis, and user presences. How do Pusher and Stream compare?

Feature

Pusher

Stream

1-to-1 chat

You must implement it on top of Pusher Channels.

Yes

Group chat

You must implement it on top of Pusher Channels.

Yes

Public and private group chat

You must implement it on top of Pusher Channels.

Yes

Chat UI components

No

Yes (iOS, Android, React, React Native, Flutter, Angular).

Push notifications

Yes, with Pusher Beams.

Yes

Message history (persisted data)

No

Yes 

Rewind/load previous messages

One message is cached for up to 30 seconds.

Yes

User presence

You must implement it on top of Pusher Channels.

Yes

Advanced chat features

None, unless you implement them.

Yes, including:

- Typing indicators
- Message status
- Unread count
- Thread replies
- Link previews
- Slash commands
- Threads
- Spam protection
- Moderation
- Image and video messages

Copy link to clipboard

Supported languages, integrations, and protocols

Can you use Pusher and Stream with the languages and frameworks that you need?

Feature

Pusher

Stream

SDKs

Pusher offers official client and server SDKs for 16 languages/frameworks.

Stream Chat offers frontend SDKs for: 

- iOS
- Android
- React and React Native
- Flutter
- Angular
- Unreal
- Unity
- Generic JavaScript

And backend SDKs for:

- NodeJS
- Python
- Ruby
- PHP
- Java
- .NET
- Java
- Go

As well as a REST API

UI widgets

No

Stream Chat offers UI components for:

- iOS
- Android
- React
- React Native
- Flutter
- Unreal
- Angular 
- Unity

3rd party integrations

Datadog and Librato for application performance monitoring.

Stream provides a Datadog integration and a REST API.

Protocols

The Pusher Channels protocol builds on top of WebSocket, falling back to HTTP streaming and polling if WebSocket is unavailable.

Stream connects chat clients via WebSocket. 

Copy link to clipboard

Security and compliance

Before you can trust your data, and your users' data, to a third-party platform you need to be sure that it meets the right security requirements.

Feature

Pusher

Stream

Encryption

Private channels support end-to-end encrpytion.

In-transit: SSL/TLS.

At rest: AES 256 encryption.

Authentication

Bearer tokens

JSON Web Tokens (JWTs)

Configurable chat roles & permissions

No

Yes

Compliance

Complies with GDPR and ISO 27001:2013.

Offers advice on how to comply with HIPAA when using Pusher.

Complies with ISO 27001:2013, SOC2 Type II, ISO 20243, HIPAA, GDPR, CCPA, DPF.

Copy link to clipboard

Pricing and support

Pusher and Stream both offer products that enable you to create chat experiences, but Stream gives you a lot more functionality out of the box. That's reflected in the pricing, with Stream's packages reflecting the fact that they offer richer functionality.

Feature

Pusher

Stream

Free plan

Free sandbox plan that offers:

- 200,000 messages per day

- 100 concurrent connections

- Support with a target response time of 48 hours

Free maker plan that offers:

- 100 monthly active users

- 25 concurrent connections

- 30 days of support 

- Limited storage

Pricing model

Pricing is based around number of messages per day and concurrent connections.

Packages include Startup, Pro, and Growth Plus.

For example, at the time of writing, the Pro plan is $99 per month and allows for 4 million messages per day with a maximum of 2,000 concurrent connections.

Pricing depends on number of monthly active users.

Start, Elevate, and Enterprise plans offer different levels of functionality, with pricing depending on the number of MAU.

For example, at the time of writing,  the Start plan is $399 per month at 10,000 MAU and $1,049 at 25,000 MAU, whereas the Pro plan adds HIPAA compliance, advanced moderation, and message translations for $599 at and $1,299 at 25,000 MAU.

Support offering

Support depends on the package you buy.

Lower priced packages offer Pusher's "Standard" level of support, which targets a 48 hour reply window, while "Premium" support targets 24 hours.

Ticket-based support is available on all paid plans, with different forms of premium support offered to Enterprise customers.

Copy link to clipboard

Pusher vs Stream: Which is the better platform?

Choosing between Pusher and Stream is a little tricky because the problems they solve exist at different layers of abstraction. Stream gives you a ready-made chat solution, whereas Pusher offers the infrastructure with which you could build your own custom chat.

So the first question is: does your problem fit Stream Chat's offering or do you need the flexibility to build something different? If you need the flexibility that Pusher offers, then Stream isn't going to be the right choice. But if you're still not sure which one offers the right solution, let's look how they compare.

Copy link to clipboard

Differences between Pusher and Stream

  • What you can build: With Pusher you can build anything that requires low latency messaging. Stream offers solutions to specific problems, such as Chat and Activity Feeds.

  • How much work you'll need to do: Stream gets you from idea to finished functionality faster than Pusher. It also reduces the amount of maintenance you'll need to do, so long as Stream fits your needs. With Pusher, you get the functionality but you have to maintain the custom solution you build on top of it.

  • Tooling and integrations: Pusher isn't a standalone solution. You'll need to integrate with a database and object storage, at the very least, to handle message history and file storage.

  • Pricing: Pusher will charge you less than Stream but you'll need to do more of your own engineering and infrastructure work if you choose Pusher.

Copy link to clipboard

Similarities between Pusher and Stream

  • Single point of failure: Pusher and Stream run your functionality in a single location. Stream's Edge network has the potential to improve latency over Pusher but neither Pusher nor Stream offers a multi-region option, making your chat application vulnerable to outages.

Copy link to clipboard

Should you choose Pusher or Stream?

For a pure chat project, where you are happy with Stream's design choices and functionality, Stream probably offers the better option. That's because Pusher will ask you to do a lot more work but without giving you much additional functionality in return. For example, Pusher can't give you guarantees over message delivery or multi-region support, meaning that you could end with a greater maintenance burden but without any noticeable improvements in user experience.

However, if you need to build a variety of low latency applications then Stream won't be able to help. However, that doesn't necessarily mean you should choose Pusher. In fact, there are several realtime platforms to choose from. In particular, Ably  gives you much greater flexibility than Stream while offering a more robust service than Pusher.

Copy link to clipboard

Build chat or any other realtime service with Ably

With Ably, you can build precisely the chat experience that meets your needs, with features such as: 

  • Online/offline status

  • Last seen indicators

  • Edit/delete messages

  • Threads

  • Emoji reactions

  • Typing indicators

  • Read receipts

  • Chat admin privileges

  • User avatar

  • Push notifications

You can find out more about our chat offering, including getting started, in our reference guide.

One of the great advantages of Ably is that it gives you the platform to build and operate any realtime functionality. That includes multiplayer collaboration, data broadcast, and IoT device communication. And Ably does that with stronger reliability and lower latency than either Pusher or Stream, thanks to:

  • Global network reach: You don't need to choose just one location. Instead, Ably distributes your workloads across a global network of edge locations.

  • <65 ms latency: Ably operates in realtime, no matter where your end users are located, thanks to ultra low latencies.

  • Strong guarantees: Ably guarantees that messages arrive in order and on time, saving you from having to build additional code to track and resend messages.

  • 99.999% uptime: Ably’s dedicated team works 24/7, so you can have peace of mind even during high-demand periods.

  • Elastic scaling: From thousands to billions of messages, Ably meets your application’s demands seamlessly.

  • A great developer experience: with SDKs targeting more than 25 languages and frameworks, integrations with common tooling, and industry leading documentation, Ably’s developer experience gives you the tools to become productive quickly.

Discover how Ably empowers you to integrate global, scalable, and efficient realtime experiences of all kinds.