Why FSharpened is needed
F# is excellent. The tooling around it has gaps that every team ends up patching privately.
The problem
F# has a superb compiler and a thoughtful language, but the surrounding developer experience is thinner than in the ecosystems it competes with. Teams rebuild the same glue again and again: analyzers, reliable hot reload, safe refactoring, reproducible state, editor-agnostic tooling and, increasingly, a safe way to let AI agents work on code.
Why not a fork?
A compiler fork is a permanent maintenance debt and splits the community. FSharpened takes the opposite stance: never fork. Everything routes through public extension points (FCS, FSharp.Analyzers.SDK, NuGet, LSP), so improvements upstream flow in and ours can flow back.
What is missing today
- One state model. Hot reload, debugging, refactoring and AI edits each invent their own checkpoints. Statepoints give them one.
- Trustworthy agent grounding. Agents guess at symbols. Semantic resolution over MCP gives them facts.
- True cross-platform parity. Tooling is too often Linux-first or Windows-only. Here every tier-1 RID is a gate.
- Safe ML in the dev loop. Models need signing, frozen judges, budgets and deterministic fallbacks to be allowed near your build.
Who it is for
F# developers who want better tooling without leaving the ecosystem, teams that want agents to help safely, and contributors who want a well-governed project with written invariants and recorded decisions.