We Love BackStage but Built CodeCargo Anyway
?1761536557115)
Open-Source Doesn't Mean Free and Easy
September 23, 2025
We love Backstage by Spotify. It was the first project that really democratized the developer experience across the enterprise. Suddenly developers, platform, and security teams had a shared place to work together. The idea of an Internal Developer Platform (IDP) felt tangible instead of theoretical. The timing was also right — toolchains have continued to get more complex and an abstraction layer was sorely needed.
In my previous role at a large tech consulting firm, I worked directly with more than 20 enterprise installations of Backstage. We kept running into the same issues. Our clients would commonly tell us:
We already have Backstage yet somehow things feel even more complicated. We only have a few people who can write the automations and that doesn't scale to our size. And most of our developers only use Backstage as a way to efficiently search for Confluence links.
At every single customer, we would see well-thought-out service catalogs (that ultimately linked back to Confluence or some other documentation tool) and beautiful UI themes. The portal looked great, but clicking the buttons didn’t actually do anything. Automation was nowhere to be found.
Why?
Because creating and maintaining automations at scale is hard. Anyone can write Terraform or create a pipeline, but ensuring that all Terraform and every single pipeline adheres to corporate security and compliance standards has historically been impossible. Most organizations started using Backstage because its free and open-source. Free simply means that you must invest your engineering talent's time to not only get it working, but also maintaining the platform and hundreds or thousands of pieces of automation.
Additional Reading: The Hidden Compliance Problem with Enterprise Pipelines
Although many of my previous customers started adopting Backstage, they quickly experienced "automation and migration fatigue." They had to adapt their current and even legacy automations to work within the platform. This took engineering time which meant less feature development and ultimately reduced revenue. Most of those enterprise Backstage installations have service catalogs that haven't been updated in months, and expose some automations that still require massive expertise to run.
Backstage became yet another IT system that didn't quite get enough investment as it required to be truly transformational. In fact, some of these systems are starting to become technical debt and add to corporate compliance including SOC2.
In my experience, this is where Backstage falls short of enterprise needs
- Service Catalog ≠ Maintainable --> every Backstage installation I've seen has a service catalog that was "up-to-date" at a point in time, but it slowly diverges from the actual software and what's deployed in production. This breaks trust and decreases usage.
- High Implementation Costs --> standing up and maintaining your own Backstage for startups works really well. However, this is difficult for enterprises with hundreds or thousands of developers. To the enterprise, its like getting a product but the batteries aren't included.
- Slow Time to Value --> get approval from your architecture review board, procure infrastructure, set up cost sharing, and work with several teams to deploy/support the application, you can finally ask your (overworked) developers to spend days/weeks populating the service catalog and enhancing their existing automations so it works within the platform.
- It's not "AI-native" --> Backstage was created before GenAI became ubiquitous. Modern tools are built on top of GenAI because when implemented well, they can turn 1 developer into 10. There's no reason why GenAI shouldn't help build and maintain your service catalog. There's no reason why GenAI shouldn't help restructure your automations to work within the platform.
Why We Built CodeCargo despite our love for Backstage
We built CodeCargo because the world does not need another developer portal. It needs a close integration to GitHub and Generative AI. It needs to automatically perform low-level tasks to free up developer time. It needs to understand organizational context to create and evaluate automations "the right way."
Instead of starting with the UI and leaving you to figure out the hard automation, we flipped the model:
- Start with Automation --> pipelines, infrastructure, security, and compliance can all be scaffolded in GitHub Actions from Day 1.
- Expose Automation through Catalog --> give developers the golden path marketplace that actually runs and allows them to code, instead of being dependent migration budgets.
- Layer Applied (Not General) AI Everywhere --> not bolted on, built natively into the workflow engine, guiding developers in real time, trained to your corporate best practices and compliance, whether it’s workflow updates or bulk migrations.
- Meet Developers Where They Are --> inside GitHub, not in yet another tool that requires yet another login. We’re a GitHub App, we make your existing investment into your future vision.
Backstage inspired us by reminding us all why we got into software.
C
CodeCargo Team
The CodeCargo team writes about GitHub workflow automation, developer productivity, and DevOps best practices.
