Swimm started as a continuous documentation platform that keeps developer docs coupled to code — docs live in the repo, reference real code snippets, and auto-sync as the code changes. The company has since expanded into agentic legacy modernization: combining proprietary deterministic static analysis, AI agents and senior engineers to understand, document and migrate large legacy systems, including COBOL, JCL and PL/I mainframe estates.
From docs to modernization
Swimm's platform builds a validated knowledge layer over a codebase that both humans and AI tools can trust. That context layer feeds coding assistants like GitHub Copilot and Claude Code (including via MCP servers), powers API-first re-architecture of embedded business logic, and de-risks large migrations such as .NET and Java upgrades or monolith-to-microservices moves. The company says it has analyzed over 100M lines of code.
Company background
Swimm was founded in 2019 in Tel Aviv by Oren Toledano, Omer Rosenbaum, Gilad Navot and Tom Ahi Dror, who previously ran the Israel Tech Challenge coding academy. It has raised $33.3M in total, including a $27.6M Series A in November 2021 led by Insight Partners and Dawn Capital.