Low-code integration platform for banking
Turn this teardown into a decision-ready prompt for ChatGPT, Claude, or your agent.
If you only have a few minutes to spare, here’s what investors, operators, and founders should know about Sandbox Banking (W17).
Sandbox Banking built Glyue, a low-code integration layer for banks, credit unions, and fintech vendors. Founded in 2015 and admitted to Y Combinator's Winter 2017 batch, the company turned a failed universal-API idea into a configurable system for moving data among core banking systems, CRMs, loan platforms, files, and event streams.[1][2]
Its exit came from solving an awkward problem well. Banking integrations require local customization, but the adapters, templates, and controls created for one institution can still become reusable assets. That combination gave Glyue both services-like depth and software-like accumulation. nCino, a longtime partner with a larger distribution base, acquired Sandbox Banking in February 2025 and later relaunched Glyue as nCino Integration Gateway.[10][12]
Ravi Balasubramanian and Skye Isard arrived at bank infrastructure from finance and software. Balasubramanian studied electrical engineering and economics at Penn and Wharton, then worked at Lehman Brothers, Barclays, UBS, and Harvard Management Company. Isard studied at Tufts, worked at MIT's AI Lab, and also spent time at Harvard Management Company.[3][4] Public sources say they lived in the same house when they started the company, but do not establish how or when they first met.
The company emerged from a different product. The founders were selling financial-advisory software to banks and credit unions when integrations repeatedly slowed deployment. Isard later told Startup Boston, “The most concise way to put it is that we encountered the problem as we were trying to solve another one.”[3] The obstruction became the opportunity.
Their first answer was too ambitious and too shallow. “We thought at first that we should build an API to encompass all of banking – a kind of universal API – and that wasn’t exactly right,” Isard said.[3] A common interface sounded elegant, but banks needed integrations tailored to their systems, controls, and workflows. After landing an initial customer, the team spent about another year turning what it had learned into Glyue, a low-code builder that preserved institution-specific configuration.
Balasubramanian framed the broader market around specialization. In an Upstart interview, he described banking as a service as a model “for banks who say, we’re actually much better off focusing on a certain portion of what we do, which is servicing relationships.”[4] Fintech partners could supply other capabilities, provided somebody made the systems work together. That dependency became Sandbox Banking's market.
Glyue sat between a financial institution's systems and the vendors it wanted to use. A team could connect a core system, CRM, loan platform, data warehouse, or fintech service without writing every integration from scratch. The platform handled web APIs, scheduled extract-transform-load jobs, file and document transfers, webhooks, and Kafka-style event pipelines.[1]
The working unit was an integration assembled from adapters and endpoints. In Glyue's Integration Builder, implementers selected connectors, mapped fields, configured external destinations, set execution and data-retention behavior, and checked compatibility with the installed engine version. Permissions, comments, and tags made the work reviewable by more than one developer.[5] Separate governance reports showed who currently held access and kept a time-stamped ledger of grants, revocations, and group changes.[6]
Read the complete post-mortem, the rebuild playbook, and the exact reasons Sandbox Banking is still worth studying now.