Build native software in a language designed for AI-native work.
Fidan is for people who are tired of choosing between Python-level ergonomics, Rust-level ceremony, and C++-level legacy drag. You get interpreter, JIT, Cranelift AOT, and LLVM AOT in one coherent toolchain, with real concurrency, direct native interop, tracing, replay, and AI-assisted explain, fix, and improve workflows that belong to the product instead of sitting beside it.
use std.async
use std.io
action main {
parallel {
task { io.print("native work") }
task { io.print("real threads") }
}
concurrent {
task {
await async.sleep(40)
io.print("cooperative scheduling")
}
}
io.print("Built for native speed.")
}
main()Fidan is built for shipping, not for ceremony.
It is built to be the language you actually want to ship native software with in 2026: compact, direct, fast, and designed like a product instead of a puzzle.
Leave runtime drag behind.
Ship native binaries, keep a smaller syntax surface than systems languages, and stop treating packaging and deployment like a second project.
Keep the performance, drop the ceremony.
Fidan stays native-first, but gets there with faster iteration, fewer concepts to juggle, and one coherent toolchain.
Modern product feel without legacy baggage.
You still target real native software, but with built-in diagnostics, formatting, LSP support, replay tooling, and a first-party package flow.
Why teams switch to Fidan
Fidan is built for people who want native performance, real deployment targets, integrated tooling, and AI-native workflows without getting dragged back into the old tradeoffs.
Native without the ceremony
Fidan keeps the language surface compact, but still gives you interpreter, JIT, Cranelift AOT, and LLVM AOT in one toolchain.
Concurrency that reads like intent
Use parallel for real threads, concurrent for structured same-thread work, and spawn/await when you want explicit handles.
Interop that does not fight you
Reach native libraries directly with @extern instead of pretending every problem belongs inside a managed sandbox.
Tooling built into the language story
Tracing, replay, packaging, diagnostics, formatting, LSP support, and compiler-grounded AI explain/fix/improve workflows are part of the product, not an afterthought.
Built-in tooling, not bolted-on AI.
The toolchain already ships with broad tests, multiple execution backends, tracing, replay, packaging workflows, and AI-assisted explain, fix, and improve paths grounded in diagnostics, inferred types, reads and writes, symbols, call graphs, and static traces. The point is not novelty. The point is a fast, coherent system that feels tight everywhere.
Dal gives Fidan a first-party package story, and the runtime stays open to native libraries through @extern so you are never boxed in.
Get from zero to native output quickly.
Use the installer for the smooth path, or pull the release archive directly if you want explicit artifact control.
iwr https://fidan.dev/install.ps1 -UseBasicParsing | iexcurl -fsSL https://fidan.dev/install.sh | shDownload the latest release from https://github.com/fidan-lang/fidan/releases/latest and add fidan to your PATH.Fidan 1.0.14
Installation
- Windows installer: download fidan_windows_bootstrap_v1.0.14.exe from this Release and run it.
- WinGet: install via winget install --id=Fidan.Fidan --source=winget
- Windows PowerShell: iwr https://fidan.dev/install.ps1 -UseBasicParsing | iex
- POSIX shell: curl -fsSL https://fidan.dev/install.sh | sh
