Step 01
Shape the workflow in preview mode
Use local-first mode to decide how ClastX should debug, build, refactor, review, and explain code before live costs exist.
ClastX docs
The guide for the ClastX application, written from the public website side.
This page should teach people what the app does, how to use it, how the workflows are structured, and how billing or upgrades fit around the application.
Step 02
Lock the coding context
Keep the product brief, target files, task definition, and success checks consistent so the assistant already feels focused.
Step 03
Turn on live services only after the workflow earns it
Once the workspace feels worth paying for, connect the OpenAI key, auth, billing, storage, and later the desktop shell.
Core modes
The app already has a clear operating model for web-development work.
Debug
Trace and fix bugs
Reproduce issues, inspect the likely layer, and ship the smallest safe patch.
reproduction pathroot-cause treefix plantest or manual check
Build
Scaffold a feature
Break a feature into user outcome, components, state, data, and implementation order.
feature planfile mapimplementation sequenceacceptance checks
Refactor
Reduce complexity safely
Find code smells, define cleaner boundaries, and move in low-risk steps.
refactor mapmigration stepsregression riskstest additions
Review
Pressure-test a change
Look for bugs, hidden regressions, missing tests, and unclear reasoning.
risk listmissing edge casestest gapsfollow-up cleanup
Explain
Understand the code path
Turn confusing code into a clear mental model with plain language and sequence.
plain-language summaryflow breakdownkey abstractionstradeoff notes
Setup surfaces
The site and app together already cover the pieces that make ClastX feel organized.
Project brief
Define the audience, stack, coding tone, and core promise in Settings so ClastX has a stable product identity.
Workspace context
Use Project context to keep the current task, files, blockers, and success definition explicit before each prompt.
Snapshots
Save local workspace sessions so important debug plans, implementation paths, and review threads are easy to reload.
Artifact pack
Generate ready prompts, execution briefs, acceptance checks, and file maps straight from the current coding mode.
Prompt library
Reusable playbooks belong to the application, but the website should explain them clearly.
Bug report
Load a prompt that turns a vague bug into a tight debug plan.
Implementation plan
Load a prompt that breaks a feature into files, state, steps, and checks.
Review checklist
Load a prompt that turns a change into a bug-risk and test-risk review.
Refactor outline
Load a prompt that maps safer boundaries and migration steps.
Explanation pass
Load a prompt that explains confusing code in plain language.
Free right now
Promotion site for ClastX and what the app does
Guides, onboarding, support, and product documentation
Pricing, billing planning, and early-access collection
The browser app shell with live AI switching and prompt flows
Saved app sessions, artifact generation, and setup context
OpenAI-ready server route and production deployment path
Turns on later
Live model usage at broader scale and tighter metering
Authentication, projects, and teams
Payments, subscriptions, entitlements, and storage
Desktop distribution and deeper repo or file-system workflows
Where to go next