Skip to main content

​
Understanding context

AI models have limited context windows — the amount of information they can process at once. As you use more of the context window, costs increase, latency increases, and response quality can degrade. Managing context effectively is essential for getting the best results from Workshop. Additionally, as projects grow and span multiple conversations, maintaining consistent understanding of what you’ve built becomes crucial for continued success.

​
The /compact command

The /compact command is the primary tool for managing long conversations. It summarizes your current conversation and automatically continues in a new child conversation, carrying forward the essential context.

​
How it works

1

Trigger compact

Type /compact in the conversation input. Optionally add focus text to guide what gets prioritized in the summary.
2

Workshop summarizes

Workshop analyzes your conversation history and creates a condensed summary of key decisions, progress, and current state.
3

New conversation starts

A new child conversation is created with the summary injected as context. You continue working seamlessly.
4

Parent becomes read-only

The original conversation is preserved as read-only history. You can review it but all new work happens in the child.

​
Usage examples

/compact
Summarizes everything and continues in a new conversation. 
/compact key decisions
Focuses the summary on important decisions made during the conversation. 
/compact focus on the authentication flow
Prioritizes the authentication-related context in the summary.

​
When to use /compact

  • Your conversation is getting long and responses are slowing down
  • Workshop warns you about approaching the context window limit
  • You want to continue a task with a fresh, focused context
  • You’ve completed a major milestone and want to start fresh for the next phase
Workshop monitors your context window usage and will proactively prompt you when you’re approaching the limit. The usage indicator at the bottom right of the app shows how much of the context window your current conversation is using.

​
Other slash commands

CommandDescriptionExample
/compact [focus]Condense and continue in new conversation/compact key decisions
/context [guidance]Update project context from conversation/context focus on API patterns
/summarize [focus]Summarize the current conversation/summarize action items

​
Project context files

Project context files give Workshop persistent knowledge about your project that carries across all conversations. These live in the .memex/ directory at your project root.

​
.memex/context.md

The primary project context file. Workshop reads this at the start of every conversation to understand your project’s architecture, current state, and conventions. How to update it:
  • Using /context: Type /context in any conversation. Workshop analyzes your project files and conversation history, then updates context.md with an intelligent summary.
  • With custom focus: Type /context focus on API patterns to steer what gets emphasized.
  • Manually: Edit .memex/context.md directly to add or correct information.
When to update:
  • After completing a major feature or milestone
  • Before starting a new conversation in an existing project
  • When Workshop seems to have forgotten important project details
  • Before making complex changes that require deep project understanding

​
.memex/rules.md

Project-specific rules that guide Workshop’s behavior. Use this for coding standards, naming conventions, framework preferences, and any other guidelines Workshop should follow consistently. 
# Project Rules

- Use TypeScript for all new files
- Follow BEM naming for CSS classes
- All API endpoints must include input validation
- Write tests for every new utility function

​
.memex/skills/ directory

Custom skills are reusable instruction sets for specialized tasks. They extend Workshop’s capabilities for project-specific workflows like deployment, testing patterns, or domain operations.

​
Custom instructions

Custom Instructions in Agent Settings apply across all your conversations and projects. Think of them as global preferences that Workshop always follows. 
Every visualization should have a transparent background and white labels.

When building web applications, prioritize clean, accessible design and responsive layouts.

Always commit working code after each major milestone.
Set them in Agent Settings (wrench icon) under Custom Instructions. See Agent Settings for details.

​
Context file precedence

Workshop loads project context using a precedence order. The first valid file found is used:
PriorityFileDescription
1 (highest).memex/context.mdWorkshop’s native context format
2.memex/rules.mdWorkshop’s rules file
3.cursor/rules/*.mdcCursor-compatible rules (all .mdc files concatenated)
4 (lowest).cursorrulesLegacy Cursor rules file
Only the highest-priority file found is loaded — they are not merged. If you have both .memex/context.md and .cursorrules, only context.md is used.

​
Context window management tips

​
Keep conversations focused

Each conversation should have a clear purpose — one feature, one bug fix, one exploration thread. Starting new conversations for different tasks keeps context clean and relevant.

​
Use /compact proactively

Don’t wait for Workshop to warn you. If a conversation has been going for a while and you’ve completed a logical milestone, compact and continue.

​
Maintain your context.md

A well-maintained context.md means Workshop starts every conversation with a solid understanding of your project. This reduces the need to re-explain architecture, conventions, and current state.

​
Provide files strategically

When giving Workshop files for context:
  • Drag and drop files into the chat (Desktop) to include their path
  • Use the file upload feature (Cloud) to add files to your project VM
  • Specify whether the file is for Workshop to read and understand (context) or to include in what it builds (content)

​
Long Context Mode

For extended sessions on large projects, enable Long Context Mode in Agent Settings. This uses an expanded 1M token context window (powered by Claude Sonnet 4.6), allowing much longer conversations before needing to compact.
Usage beyond the standard 200k context window may result in higher credit usage per turn, in alignment with Claude Sonnet 4.6 long context pricing.