Superglue Wants to Kill Every Brittle SQL Script Holding Your Company Together

The Macro: The Glue Code Problem Is Eating Engineering Teams Alive

Every company with more than about 50 employees has the same dirty secret: critical business processes run on integration code that nobody fully understands. SQL scripts that sync customer data between Salesforce and the data warehouse. Cron jobs that pull invoices from one system and push them into another. FTP uploads that someone set up in 2019 and nobody has touched since because the person who wrote it left two years ago.

This is glue code. It’s the connective tissue between enterprise systems, and it is almost universally terrible. Not because the people who wrote it were bad engineers, but because integrations are treated as afterthoughts. The real product gets the architecture reviews and the code quality standards. The integration that moves data between the real product and the accounting system gets written in a sprint, tested once, and forgotten.

The traditional integration market is enormous and fractured. MuleSoft (now Salesforce) is the enterprise incumbent, and it costs accordingly. Workato, Tray.io, and Merge target mid-market with visual workflow builders. Fivetran and Airbyte handle data pipeline integrations specifically. Zapier and Make cover the low end. None of these solve the fundamental problem, which is that integrations break constantly and nobody wants to maintain them.

AI changes the calculus in two ways. First, AI can read and understand legacy integration code, which means it can reverse-engineer what a script does even when there’s no documentation. Second, AI can monitor integrations and fix them when they break, without waiting for a human to notice the silent failure at 3 AM.

The company that can actually deliver self-healing integrations will save enterprises millions of hours of engineering time annually. That’s not an exaggeration. Gartner estimates that integration work consumes 30-40% of enterprise IT budgets.

The Micro: Open Source, Self-Healing, and SOC 2 Ready

Superglue positions itself as an “agentic integration platform.” The product takes legacy integration code, whether it’s SQL scripts, FTP pipelines, or custom connectors, and converts it into production-grade API integrations with versioning, monitoring, and auto-repair.

The auto-repair claim is the interesting one. When an integration breaks (an API endpoint changes, a field gets renamed, a schema migrates), superglue is supposed to detect the failure and fix the integration automatically. If that works reliably, it eliminates the most painful part of integration maintenance: the 2 AM pages, the data discrepancies that nobody notices for three days, the cascading failures when one system change breaks five downstream processes.

Adina Goerres and Stefan Faistenauer are the founders. Adina previously worked in GTM at Scalable Capital and on commercial projects at Farfetch. She has a Master’s from TUM and CDTM. Stefan was a software engineer and Google APM with a Stanford Master’s. They’re a four-person team out of San Francisco, part of YC’s Winter 2025 batch with Tom Blomfield as their partner.

The Google APM background is relevant. Google’s Associate Product Manager program is notoriously selective and produces people who understand both product thinking and engineering constraints. Stefan has seen enterprise infrastructure at Google scale, which is useful context when you’re building a tool that’s supposed to handle enterprise complexity.

The product is open source on GitHub (github.com/superglue-ai/superglue), which is a smart go-to-market motion for developer-facing infrastructure. Enterprises can evaluate it without a sales call, and open source creates trust in a category where vendors have historically locked customers into proprietary platforms. They also offer self-hosting, auto-scaling, white-label options, and claim SOC 2 Type II and ISO 27001 compliance.

The customer list apparently includes names like Netlify, which suggests they’re getting traction with companies that have real integration complexity. They claim a 70% reduction in migration effort, which is a big number if it holds up in practice.

The Verdict

I think superglue is attacking a problem that every engineering leader complains about but few startups have seriously tried to solve. The integration layer at most companies is a mess because nobody prioritizes it, and nobody prioritizes it because the tools for managing it have historically been either too expensive or too limited.

The open-source approach is smart and the technical team has the right background. The risk is that “self-healing integrations” is one of those features that’s incredibly hard to make reliable. An integration that fixes itself 90% of the time and silently corrupts data the other 10% is worse than no automation at all. Trust is everything in this space.

At 30 days, I’d want to see how many legacy integrations customers are actually migrating onto the platform, and what the failure rate looks like. At 60 days, the auto-repair feature needs data: how often does it fire, and what’s the success rate? At 90 days, the question is whether enterprises are using superglue for new integrations too, not just migrating old ones. If it becomes the default integration layer for new projects, the growth flywheel kicks in. If it stays a migration tool, the market is large but one-time. The problem is real, the team is credible, and the open-source distribution gives them a genuine advantage over closed-source competitors. This one is worth watching.