Spotify may be synonymous with music streaming, but it’s also got a wildly popular developer-tool side-hustle called “Backstage.”
Backstage is an open source project that helps companies build their own internal developer portals: a catalog of their developer tools along with quick visualizations of the work the tools have done, and other metrics. But like many open-source projects, Backstage is a build-it-yourself option.
Israeli startup Port has been gaining big-name customers like GitHub, British Telecom and LG with a proprietary Backstage competitor: a dev tool portal that’s also now been geared to manage AI agents.
On Thursday, Port, founded in 2022, said it raised a fresh $100 million Series C round led by General Atlantic, with participation from Accel, Bessemer Venture Partners and Team8. The round values Port at $800 million and brings its total funding to date to $158 million. This Series C follows the company’s $35 million Series B led by Accel and Bessemer, announced in May.
Of all the industries that LLM-based tech has infiltrated, coding is where it has the deepest roots. So, not surprisingly, developers are also on the cutting edge of building and adopting agents that can automate entire repeated processes — work far beyond asking AI to write some code.
But the problem here, according to Port co-founder and CEO Zohar Einy, it’s the wild west right now for such devtool agents at companies: finding them, sharing them, ensuring their work follows company standards and so on.
Developers “want to take AI beyond just coding. They want it to resolve incidents, resolve security issues. They want it to take care of the release management,” Einy told TechCrunch.
Techcrunch event
San Francisco
|
October 13-15, 2026
But if agents are connected to all kinds of different tools and data sources, if the data is scattered among them, if they have no way to collaborate, and have no corporate standards and guardrails, “it creates chaos,” his product pitch goes.
Port therefore offers more than just a catalog of dev and agent tools (although it does offer that). It supplies a layer of orchestration with features that measure agent performance and add a human-in-the-loop, as desired, to approval processes.
A feature called “context lake” defines the data sources, context memory, and guardrails for agents. “It’s where you manage what agents ‘need to know’ to do their job safely and correctly,” Einy explained.
In addition to using Port to catalog agents devs have already created using other tools, they can use Port to create new agents. Plus, Port also offers a few of its own ready-made agents, which can do things like resolving helpdesk tickets and dealing with provisioning.
Einy describes his product as handling the other 90% of what software programmers do that isn’t writing code. “It gives the engineers a user interface to control the agent, to iterate with the agent, to approve what it does that is not coding, that is all the 90%.”
With its giant new war chest of cash, big name customers and tier-one VCs, Port looks like an agentic management startup to watch. But to say it faces competition is an understatement.
The entire category of agentic management and orchestration is flooded with hopefuls, from big tech companies to startups, and they’re all coming at the various new problems in the space from different angles. A few of these include LangChain, UiPath, Cortex and more.
