Completed
Fully subscribed and closed. Thank you to everyone who backed this round.
CodeHarborThis project has been verified
AI pair-programmer that actually understands 40-year-old COBOL.
$850,000 raised so far, confirmed by both parties (100%)
Details
- Category
- Developer Tools
- IP type
- Software
- Stage
- Scaling
- Ask type
- Fixed amount
- Listed
- Sep 16, 2026
- Views
- 1
Description
CodeHarbor is an AI coding assistant purpose-built for mainframe modernization: it reads COBOL, PL/I, and JCL alongside a codebase's actual runtime behavior, then proposes safe, incremental migrations to modern languages without breaking decades of untested business logic. Banks and insurers use it to cut migration timelines from years to months.
Problem
Trillions of dollars in financial infrastructure still runs on COBOL, and the engineers who understand it are retiring faster than anyone is replacing them.
Solution
A static-and-dynamic analysis engine plus an LLM fine-tuned specifically on mainframe languages, producing verified, test-backed migration diffs instead of risky rewrites.
Target market
Banks, insurers, and government agencies with large legacy COBOL estates.
Business model
Enterprise SaaS, priced per line of code under management, with an annual support contract.
Investments
James Whitfield invested $350,000
8% equity, board observer seat
Yuki Tanaka invested $250,000
6% equity
Helena Voss invested $250,000
6% equity
Related projects
Discussion (2)
James Whitfield
Ran this against a real core-banking COBOL estate in diligence, and the migration diffs were genuinely usable, not just plausible-looking.
Yuki Tanaka
The kind of unglamorous problem that's actually worth a lot of money to solve well.
Priya NakamuraVerified account
Innovator·🇯🇵 Japan
Ex-infra lead, now building developer tools full time.