Sisense's Compose SDK marks a shift toward a headless architecture. Launched in 2023, it gives developers more control over the front-end design of embedded analytics.
If you’re evaluating the Compose SDK, you’re likely interested in building a flexible analytics layer on top of your application code. But beyond customization, you should also look at performance, security, integration complexity, and cost.
Let’s break down what Compose SDK is, how it compares to Embeddable, and which one might be a better fit for your team.
What is the Compose SDK?
Compose SDK is a software development kit that allows developers to build analytics features directly into their application code. Instead of relying on predefined dashboards, you can create Sisense queries on the fly and display results using your UI components.
Built for modern frontend workflows, Compose SDK supports common UI frameworks like React and Vue. It enables developers to render existing widgets or build custom visualizations, drawing on Sisense platform capabilities through SDK packages and APIs.
This flexibility does come with some setup requirements. You still need a working Sisense instance and access to Sisense’s backend infrastructure. The SDK acts as a bridge between your client side libraries and Sisense’s powerful analytics platform.
To support implementation, Sisense provides documentation, CLI tools, and code samples that guide you through tasks like query composition, data binding, and rendering visuals. You’ll also need to configure authentication methods, such as an API token, to securely connect your frontend to the backend.
Compose SDK can be an interesting option for teams upgrading existing Sisense projects or starting fresh with embedded analytics inside customer-facing apps.
Ultimately, the Compose SDK is a plug-in to an architecture that was originally built for internal BI – where it doesn’t matter how it looks or how fast it loads – which was then modified to support embedded analytics. As a result it retains some of the artefacts of its history that lead to complexity and feature bloat.
If you’re looking for more flexibility and control over your customer-facing dashboards and analytics, it’s worth exploring the Compose SDK alongside tools that were built from the ground up with the headless architecture that gives you this extra control.
Embeddable: truly headless embedded analytics
Embeddable is a set of flexible, extendable building blocks for engineering and product teams to deliver customer-facing analytics projects. The architecture was built from day 1 as a true headless embedded analytics tool - meaning there are no workarounds or artefacts of previous evolutions. It just works.
With Embeddable you can import any React charts, use the charts provided, or extend without restrictions. It will feel like it’s custom built, but won’t take endless engineering time to complete.
It uses a simple Web Component or React SDK that plugs directly into your frontend. You connect Embeddable to a queryable data source, and it handles the backend processing, security, and caching.
Embeddable is ideal if you want to build analytics that feel like a natural extension of your app. There are no widgets required unless you choose to use them, and no constraints on layout or theme.
Performance is a priority. Embeddable uses Cube for semantic modeling and fast caching, and it also connects to the fastest modern databases, making it a strong fit for real-time analytics or high-volume user environments.
Key differences between Compose SDK and Embeddable
The two options may seem similar at a glance, but they're different in many ways that concern your dev team and your end-users.
1. Architecture
Compose SDK is a plug-in that extends the Sisense platform to enable more flexibility in the UX and UI. However, it relies on a deployed Sisense instance for data access and processing and comes with all of the logic and requirements of an architecture that was built initially for internal BI – where the requirements for customisation and speed of loading are not as high as for customer-facing applications. It also has all of the features of a BI tool, which can feel heavy and bloated.
Embeddable is a standalone tool designed from the ground up for embedded analytics with no legacy BI platform requirements. It provides a library of charts to get started with, as code, allowing you to infinitely extend and modify the components so they feel native in your platform. It’s built with developers in mind, meaning architectural decisions were made to give you the customization freedom of a custom build, with the quick time-to-market of an off-the-shelf tool.
“Embeddable has enabled us to improve the experience for our users in a way and timespan that we wouldn't have achieved otherwise.” Bart Lammers, CoFounder & CTO at Eyk
2. Developer experience
Compose SDK offers strong control, but within the bounds of Sisense’s APIs and data models. It uses client side libraries and supports query composition from within your app.
While both tools can embed in any frontend, Compose SDK does allow you to build your charts in Vue whereas Embeddable’s SDK only enables you to import/extend charts in React.
Embeddable is designed to integrate natively into the development processes of modern software applications. It works with your CI/CD and embeds either natively as react or via a web component (no iframes) - meaning you can hook it up to your existing monitoring and you can pass variables in and out of the dashboards (e.g. configuring dark/light mode when it’s set at the global app level).
"The developer experience with Embeddable is excellent. The platform is designed with developers in mind, offering tools and features that streamline the development process. It supports rapid development cycles, enabling us to focus on creating and refining our applications without unnecessary complications." - Pierre Vaidie, Co-founder & Head of Engineering at Pledge
3. UX/UI customization
Now that Sisense has released the Compose SDK, both tools now offer a lot of flexibility and control in the end-user experience, enabling you to deliver an experience that feels native in your application.
With Compose SDK, you can render existing widgets or create custom ones in Vue or React. You can also style and position components as desired.
Similarly, Embeddable gives you full UI freedom from the start. You can use the charts provided, extend them infinitely, apply your own design system and even import your own charting library and interactions if you need a custom chart type.
4. Performance and scalability
Sisense’s backend is powerful but top performance is only offered as part of the enterprise packages, and may require tuning and infrastructure management. The Compose SDK lets you selectively load only the required widgets, which can help to improve performance.
Embeddable is optimized to enable you to load dashboards in sub-second response times, just like every other page in your application. Its data service and caching layers are provided as standard, so you won’t need an enterprise plan to get the best out of it. It connects to the latest real-time databases like Clickhouse, StarTree and Pinot. You can run the cache as part of the Embeddable service or connect to your own instance of Cube so that the cached data is stored on your own environment if you want it to be.
5. Integration complexity
Using a headless embedded analytics tool gives you much more power over the UX and UI that your customers will see, but access to code will increase reliance on developers to a degree.
The great thing about headless tools is that most of the coding is done during setup, and then the charts and dashboards can be managed almost entirely in a no-code dashboard management platform. This is true for both tools, however there are some differences.
On top of styling/defining your charts:
Sisense’s Compose SDK requires setup of the Sisense UI, API token configuration, and data model preparation. Implementing row-level security for multi-tenant applications is also more complex (see below) and its embedding options present varying degrees of complexity (iframe or SDK embed).
Embeddable requires you to connect your database, define your data models and then embed as a web component or as a native React embed - it requires no concept of your users, instead provides developer-defined, code-level RLS to provide simplicity and flexibility to ensure users see only the data they are supposed to. The dashboard builder is also simple and flexible, enabling you to easily create interactions between charts and filters via variables.
“The no-code builder is very intuitive to use, it contains what is needed to create awesome visualisations and dashboards” - Rodel van Rooijen, Founding Engineer at Solvimon.
6. Pricing
To use the Compose SDK, you’ll need a Sisense license. This may be in addition to a license you already have. It’s quote-based and users report that it’s typically based on usage or the number of users. You can read our guide on Sisense’s pricing here.
Embeddable offers flat pricing, which is more predictable for teams embedding analytics into SaaS products, especially where userbases and usage can fluctuate. It’s also quote-based, but not based on usage or users, so you’ll pay the same price every month, no matter how many users you have and how much data they use.
If you have a project you’d like to get a quote for, you can fill in this form and speak to one of the team.
7. Row-level security (RLS)
If you’re building a multi-tenant SaaS app, Embeddable’s RLS is typically a better fit. Sisense works well in BI team environments but adds friction when trying to externalize analytics at scale.
Sisense handles RLS using a data security policy engine and user/group filtering, or both:
- User & Group Filtering: RLS is applied by associating users or groups with filters on specific data models or dashboards. Filters are written in SQL-like expressions, e.g., Region = "East".
- Data Security Policies: Admins define data security rules at the Elasticube or Live Model level. These rules are evaluated server-side during query execution.
- Multiple Mechanisms:
- Elasticubes: Use in-memory models and static RLS rules.
- Live Models: Pass security filters as part of live queries (to Snowflake, BigQuery, etc.).
- Session-Based Logic: Can reference the logged-in user’s attributes (user_name, group_name, etc.) to apply dynamic filters.
Here are some of the limitations:
- Complex to manage at scale—especially with many users/tenants.
- Tied tightly to Sisense’s own data models; not headless.
- Not easily programmable via APIs—requires dashboard-centric configuration.
Embeddable uses developer-defined, code-level RLS, typically based on JWTs (JSON Web Tokens). It’s easy to implement multitenant logic or per-customer filtering with code, configuration, or middleware.
Unlike Sisense, Embeddable is backend-first and gives full control to developers over how RLS is enforced on every query. You can dynamically generate queries based on user/session data. RLS is enforced through backend code or query templates. JWTs include attributes (e.g., user_id, org_id, role) that get injected into the query context.
Here are some of the benefits:
- Fully decoupled from UI/dashboard tools.
- Easy to version and deploy RLS logic as code.
- More scalable for SaaS use cases with many tenants.
8. Use cases
Embeddable has run many headless embedded analytics projects with many customers who have shared their experience. It’s been used for projects at small companies, scale-ups and enterprises - and specific examples with quotes from those using it are available as case studies. Note how each of the examples is completely unique in their styling and functionality. This demonstrates the power of true headless embedded analytics.
Compose SDK is a relatively new offering, hence user feedback and use cases are sparse. After a good search on the internet, and interrogating ChatGPT, there were no publicly-available use cases or quotes to the positive or negative for the Compose SDK. We hope this will evolve over time, and will share examples here once they become available.
Which one should you choose?
If your team already uses Sisense and wants to extend its capabilities, Compose SDK is a natural progression. You’ll be able to leverage the existing platform while building more flexible, code-driven interfaces.
If you're starting from scratch and want to embed analytics into a client-facing product, Embeddable offers speed, simplicity, and a native feel. It’s built from the ground up for product and engineering teams who want to build custom analytics experiences without the overhead of a custom build, and without compromising by embedding a clunky BI tool that wasn’t build for the job.
To summarize:
- Choose Compose SDK if you're expanding an existing Sisense instance or want tight control over data access within Sisense.
- Choose Embeddable if you want to quickly build analytics with a modern, customizable UI and lower overhead.
If you’re interested in testing fully headless embedded analytics, Embeddable currently offers a supported PoC to enable you to assess it for your project without a large upfront commitment. Find out more about how to use Embeddable for your project.
FAQs
What is a headless architecture?
A headless architecture decouples the front end from the backend. The front end uses APIs to pull data and render it however you want rather than presenting a third-party vendor’s UI to customers in your application.
What is headless embedded analytics?
It’s an approach to embedding analytics that enables you to extend, modify and import charts and components in code. This gives you total control over UX, UI and layout - whilst saving you engineering time vs. a fully custom build.
How is Embeddable different from Compose SDK?
Embeddable is a dedicated, headless tool for building customer-facing dashboards and analytics with full UX/UI control. Compose SDK is an SDK layer on top of Sisense’s analytics platform that gives existing users some additional flexibility.
