Undo builds deterministic program recording technology that captures complete execution history into self-contained recordings to give AI agents runtime context for diagnosing software failures across development, test, and production. The company says this runtime context improves AI model root-cause identification from 38% to 92% on a range of complex bugs and that customers report completing root-cause analysis up to 100x faster. Undo also reports that using its recordings reduces token usage when models solve problems and that its technology helps surface issues living in runtime state not captured by logs. Palo Alto Networks is cited as a customer that uses Undo to find root causes in minutes in large codebases. Undo plans to scale product development, customer support, and go-to-market teams across the US and Europe following the funding. The company was founded in 2012 and is based in Cambridge.

Total raised

$55M

Funding rounds

5

Latest round

Equity

Latest activity

Jun 2026