A hosted developer portal with a software catalogue, scorecards and self-service workflows is where I'd start for a small or mid-sized platform team, because it fills faster than an open-source framework you run and maintain yourself. Choose a platform orchestrator instead when your bottleneck is provisioning infrastructure rather than knowing who owns which service.
What a wrong platform choice costs
Pick the wrong internal developer platform and you won't find out in the demo. You'll find out a year later, when the catalogue is half-filled, the scorecards measure things nobody looks at, and the platform team spends its week patching plugins instead of building the golden paths developers wanted. Then you migrate, which means re-importing every service and re-teaching every team where to look. That's the expensive mistake, and it's avoidable.
I'd put most of the risk on buying the wrong type, not the wrong brand. The category splits three ways. Open-source frameworks give you the building blocks and leave the running to you. Hosted portals give you a catalogue, scorecards and workflows as a service, sometimes built on the same open-source base. Platform orchestrators and control planes go further and provision infrastructure, so they sit closer to your cloud. The questions further down walk through how to choose between them.
The market is also shifting. Atlassian says its Compass scorecards and catalog functionality will move into DX, and agentic features are appearing across these products, from maintenance agents to self-service workflows run by AI. Neither changes the core decision, but both are good reasons to ask each vendor where its roadmap is heading. If AI tooling for your developers is on your list too, there's a wider round-up of AI tools for developers.
Internal developer platforms and portals to compare
Backstage
The open-source base of this category: a developer portal framework with a software catalogue, software templates and TechDocs.
Port
Maps your engineering environment into a catalogue, then runs self-service actions and agentic workflows on top of it.
Cortex
An engineering operations platform that combines a service catalogue with scorecards and reliability tracking.
OpsLevel
An internal developer portal built around a context catalogue, workflows and an agent that handles maintenance chores.
Harness Internal Developer Portal
Harness's developer portal, with self-service workflows so developers can get what they need without filing a ticket.
Roadie
Hosted Backstage for teams that want the open-source portal without running and upgrading it themselves.
Humanitec
A platform orchestrator with a portal, for teams building a full internal developer platform, hosted or self-hosted.
Qovery
An infrastructure control plane that lets developers run and govern projects on any cloud from one place.
Cycloid
Combines an internal developer platform and portal with self-service and an inventory of your cloud assets.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Kratix
An open-source framework for platform teams who want to build their own platform capabilities and deliver them consistently.
Before you book a demo
Write down the one problem you're buying for, in a sentence your developers would recognise: "nobody knows who owns the billing service" or "a new environment takes a week". Take that sentence into every demo and make the vendor solve it on screen with something close to your setup. If the portal is all you need, there's a narrower list of developer portal tools, and the answers below cover the rest.
Do you make internal developer platforms or developer portals? Tool makers can ask to be considered for a featured place on this list.
Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.
Frequently asked questions
What's the difference between an internal developer portal and an internal developer platform?
The portal is the front door: a catalogue of services, owners, docs and templates, plus scorecards that show whether each service meets your standards. The platform is the machinery behind it that provisions environments, wires up infrastructure and enforces how things get deployed. Some products are mainly portals and read from what you already run. Others are orchestrators or control planes that create the infrastructure themselves, and some sell both. When a vendor says "platform", ask which half you're buying. If your developers keep asking who owns a service or where its runbook lives, you need a portal. If they're waiting days for an environment, you need the orchestration side, and a prettier catalogue won't fix it.
Should we self-host an open-source portal or pay for a hosted one?
The open-source route gives you the base pieces, a software catalogue, templates and docs, for no licence fee, and full control over plugins and data. What it costs you is engineering time: someone has to run it, upgrade it, write and maintain plugins and keep it healthy, and that someone is usually on the platform team you were trying to free up. A hosted portal, including hosted versions of the open-source framework, moves that work to the vendor. I'd only self-host if you have a platform team with the appetite to treat the portal as a product it owns for years. Otherwise, pay for hosting and spend your engineers on golden paths and templates, the part that helps developers.
How are internal developer platforms priced?
There's no single model, so compare carefully. Some charge per developer or per seat each month, which is easy to forecast and grows with headcount. Others charge a flat monthly fee for the platform with a set number of users included, which suits a larger engineering team and stings for a small one. Free tiers exist for small teams on some hosted portals, and open-source frameworks have no licence fee at all, only the cost of running them. Prices sit on each vendor's pricing page, though enterprise tiers often need a sales call. When you compare, count who needs a seat: every developer who views the catalogue, or only the people who edit it? That one question can change the total more than the headline rate.
What should go into the software catalogue first?
Start with ownership. For every service, record the owning team, the on-call rota or contact, the repository, and where its docs live. That alone answers the question developers ask most often in chat. Next, add the dependencies between services, so an incident responder can see what else breaks. Only then add scorecards, and keep the first ones simple: does the service have an owner, a runbook and an up-to-date dependency list? Resist importing everything on day one. A catalogue that's half-wrong is trusted by nobody, and once developers stop trusting it they stop updating it. Fill one domain or one team's services well, prove it saves time, and expand from there.
How do we trial a portal before signing a contract?
Give it a fortnight and one real team. Connect your source control and one cloud account, import that team's services, and set three goals: every service has an owner, one self-service template creates a new service end to end, and one scorecard runs against real data. Then ask the team three questions at the end. Did they find anything faster? Did the template save them a ticket? Would they notice if the portal disappeared tomorrow? Keep a note of how many hours your platform engineers spent on setup, because that's the hidden cost in the comparison. If a vendor can't support a trial with your real systems, treat that as information about how the rest of the relationship will go.
Which type suits a small platform team?
If you're two or three platform engineers supporting a few dozen developers, I'd buy a hosted portal with a free or seat-based entry tier and resist building anything. Your scarce resource is time, and running an open-source framework yourself will eat it. Look for strong templates and self-service workflows, because those cut tickets. A platform orchestrator with a flat monthly fee only makes sense once provisioning is the real queue. For larger organisations with dedicated platform teams, the calculation changes: self-running an open-source framework or a framework for building platform capabilities becomes reasonable, because you have the people to own it and the scale to justify the control.


