LibraryFrontendAI CodingOpen Sourceeffect-tstypescriptagentstesting
Foldkit
TypeScript frontend framework built on Effect that uses The Elm Architecture. One immutable Model, typed Messages, a single update function, and side effects returned as Commands.
Added October 18, 2026Devin Jameson (@devinjameson)
Scaffold with:
bash
npx create-foldkit-app@latestThe Model is defined with Effect Schema, and every change goes through update, which returns the next Model plus any Commands. Commands only describe a side effect, and the runtime runs it. Pure updates make testing easy. Story tests read like Given/When/Then: given(model), then message(...), then assert on the Commands and the resulting Model. Scene tests exercise the view through accessible locators without a DOM. DevTools logs every Message and can rewind the Model, and a DevTools MCP server gives agents the same view of a running app: they can inspect state, query Message history and dispatch Messages. MIT, pre-1.0, so minor releases can still break things.
It's the F in the CAFE stack (Cloudflare, Alchemy, Foldkit, Effect). The idea is that with a fixed loop of Messages, update and Commands, a coding agent has less architecture to invent. Alchemy covers the infrastructure as TypeScript ("Infrastructure as Effects") and has a Cloudflare.Website.Foldkit resource that deploys an app without a Wrangler config. The reference app is Effect-TS/cafe, a chat app where agents write a failing test first, get it passing locally with alchemy dev, then rerun it against real Cloudflare. Cloudflare also has an official Foldkit guide for Workers that covers SPA, SSG and SSR.
Repo: foldkit/foldkit
Related
- T3 CodeTheo's open-source control plane for coding agents. Drive Claude Code, Codex, Cursor, OpenCode, Pi and ACP agents on your own machine with your existing subscriptions, from a desktop, web or mobile app.
- OpenCodeOpen-source AI coding agent from Anomaly (formerly SST) that runs in the terminal, as a desktop app, or as an IDE extension, and works with 75+ LLM providers including local models.
- Pi DurableDurable harness for long-running agents from the team behind the Pi coding agent. It checkpoints every step so a run survives crashes and evictions, and it now runs inside a Cloudflare Durable Object through the Agents SDK's PiHarness.
- e2eTesterArmy's open-source TypeScript test runner for web and mobile apps, where each test can be fully deterministic, fully agent-driven from natural-language goals, or a mix of both.