Local-first static site workbench

Ship evidence,
not hope.

CodeForge opens a static site project in your browser and gives you the tools to change it safely: find it everywhere and fix it once, run quality gates, checkpoint before you move, review what the browser pulls, and prove every route and form before release. Nothing leaves your machine.

8connected workbench views
0network calls during analysis
100%of checks run on local source

How it works

Change a site with proof

Open the project

Load your static site into the workspace. Files stay on your machine - the studio reads and writes locally.

Find it everywhere, fix it once

Project-wide search shows every hit in context; a fix applies across files with the hits in front of you, not behind you.

Gate the release

Quality checks, a dependency review, and a route crawl produce the evidence. Package the release from what you proved, not what you hope.

Workspace

The project, in front of you

Local files

Your machine, your files

The workspace opens a project folder in the browser. Source never transits a server.

Edit

Changes in place

Edit files directly, with the rest of the workbench - search, checks, history - watching the same source.

Commands

Command palette

Every action is reachable from the commands view, so the workbench is fast once you know it.

Quality gates

Ship evidence

Gates

Checks with verdicts

Run the gates and get pass/fail with the failing detail attached. A green release means the checks ran, not that someone felt good.

Release

Package what you proved

The release view assembles the evidence - checks, diffs, route crawl - into the package you ship.

Checkpoints

See exactly what changed

Checkpoint the project before a risky move, then diff against it after. The history view shows exactly what changed between any two points - file by file.

Dependency review

Know what the browser pulls

A static scan of your source for remote scripts, styles, images, links, and module imports. This is analysis of your files, not a package install and not a network request - the list is what a visitor's browser would fetch.

Route graph

Prove every route and form

Crawl

Links resolved

Internal links are followed and resolved across your local HTML files.

Fragments

Anchors verified

Fragment targets are checked, so a #docs link never lands nowhere.

Hygiene

Orphans and titles

Orphan pages, duplicate titles, and unlabeled form controls get flagged before visitors find them.

Local-first

Analysis without a network

Every scan, check, and crawl runs against local source in your browser. The dependency review lists remote resources by reading your files - it never fetches them. What you build here is yours; what you prove here is reproducible.