Feature Demonstrations

Discover the most advanced Gluon v2 features that automate your AI workflow and save time.

🎯

Auto Select - Automatic File Selection

Let AI select appropriate files based on task description

How does it work?

  1. Describe task in Quick Task field (e.g. "Fix login bug")
  2. Click Auto Select button
  3. Gluon sends project structure + description to AI
  4. AI analyzes and suggests files
  5. Overlay with suggestions appears
  6. Click "Apply" to automatically select all

❌ Before Auto Select

Manual file tree browsing, guessing which files are needed, wasting 1-3 minutes per task

✅ After Auto Select

AI automatically selects all needed files in seconds, zero guessing, instant work start

Auto Select Task Input

Step 1: Describe task

AI Studio Output

Step 2: AI suggests files

💎 Benefits

  • ⚡ Time saving: 1-3 minutes saved per task
  • 🎯 Precision: AI selects files more accurately than human
  • 🧠 Context understanding: AI understands file dependencies
  • 💡 Learning: See which files AI considers relevant
💟

Context Save - Saving Context

Never lose perfect file configuration

How does it work?

  1. Select files for specific task
  2. Click Context Save button
  3. AI generates summary of what you're doing
  4. Context saves with name, files and environment

❌ Without Context Save

Working on a bug for 2 hours. AI limit ends. Next day: "What was I doing? Which files? Where was I?"
Another 30 minutes to recreate context...

✅ With Context Save

Click Context Save. AI saves: what you did, which files, what you fixed, what's left.
Next day: click tile → new AI session has FULL context. Start in 5 seconds.

Context Save AI Output

AI generates name and summary

Context Save Example List

List of saved contexts

✹

Prompt Generator

Optimize AI communication

How does it work?

  1. Select files and describe task generally
  2. Click Prompt Generator
  3. AI analyzes selected code and task
  4. Generates optimized, detailed prompt
  5. Gluon prepares ready context package

❌ Your Prompt

"Napraw błąd logowania"
→

✅ Generated Prompt

"Based on the provided files (auth.js, login.component.ts, api.service.ts),
analyze the login flow and identify the bug causing authentication
failures. Specifically:
1. Check token validation in auth.js line 45-60
2. Verify API endpoint call in api.service.ts
3. Examine error handling in login.component.ts
4. Suggest fix with code examples"
Prompt Generator Output

AI generates optimized prompt

💎 Benefits

  • 🎯 Better answers: Detailed prompt = more precise AI responses
  • ⚡ Token saving: AI understands context immediately without follow-ups
  • 📚 Learning: See how to write good prompts
  • 🚀 Speed: No "AI didn't understand → clarify" iterations
⚙

Template Configuration

Precisely control how AI processes your code

What are templates?

Templates define AI function behavior - Auto Select, Context Handoff and Prompt Handoff. Each template is a set of instructions on how AI should analyze and format results.

❌ Without Custom Template

AI uses default instructions. Results are generic, no project-specific details.

✅ With Custom Template

AI works by your rules. Understands project architecture, uses proper formats, generates precise results.

🎯 Configuration Fields

Basic:

  • Template Name: Template name (e.g., "React Spec Generator")
  • Template Type:
    • Auto Select - file selection
    • Context Handoff - context transfer between sessions
    • Prompt Handoff - prompt generation
  • Role / Behavior Definition: Defines AI role (e.g., "You are a React architect, specializing in Next.js and TypeScript")

📝 Dynamic Fields - Context Handoff:

Context Handoff allows transferring full work context to a new AI session. Define how AI should summarize your work so far.

summary - Work history summary

Purpose: Defines how AI should create a chronological summary of the entire work session - what you did, in what order, what decisions you made.

How to use: Use instructions like "Create chronological history of all actions in this session, including decisions made and their reasons." When you move to a new AI session, you'll get full context of what happened.

Example value: "Present detailed work chronology: what I did step by step, which files I modified, what problems I encountered and how I solved them."

solved_problems - Solved problems

Purpose: Specifies how AI should describe problems you've already solved in this session.

How to use: Ask AI to list each bug/problem with solution description. This prevents duplicate work - new AI session knows what's already working.

Example value: "List of problems I fixed: problem name, cause, applied solution, files I changed."

current_problem - Current status

Purpose: Describes what you're currently working on and what's left to do.

How to use: AI in new session immediately knows where to continue. Instead of "what was I doing?" you have a ready starting point.

Example value: "Describe current problem I'm working on: what I'm trying to achieve, what I've already done, what blocks further progress."

key_insights - Key insights

Purpose: Gathers important technical discoveries from session - how code works, hidden dependencies, architecture.

How to use: AI saves project knowledge gained during work. In new session you use this knowledge instead of rediscovering everything.

Example value: "List of key technical discoveries: how architecture works, what dependencies exist between modules, what patterns are used."

📝 Dynamic Fields - Prompt Handoff:

Prompt Handoff analyzes your code and general task description, then generates detailed, optimized prompt for the next AI session.

task_description - Task goal description

Purpose: Defines how AI should describe the main task goal - what exactly you want to achieve.

How to use: Transform general "fix bug" into detailed description with context. AI in new session immediately understands WHY you're doing this task and WHAT is the expected result.

Example value: "Based on selected files and my task description, describe in detail the goal: what needs to be implemented, what problem to solve, what's the business context."

implementation_steps - Implementation plan

Purpose: Specifies implementation plan format - how AI should break down task into specific steps.

How to use: Instead of guessing "what to do next", you have a ready list of atomic steps. Each step is a concrete action you can execute.

Example value: "Create implementation plan: atomic steps (one file/one function), in logical order, with dependencies between steps."

technologies - Technology stack

Purpose: Describes how AI should list technologies, libraries, and tools used in the task.

How to use: AI in new session knows what technologies are involved. Thanks to this, it suggests solutions compatible with project stack, doesn't propose tools you don't use.

Example value: "List all technologies used in selected files: framework, libraries, build tools, versions if relevant."

architecture - Architecture and flow

Purpose: Defines how AI should describe application architecture and data flow in task context.

How to use: AI understands how components work together. Instead of "change file X", AI sees "change file X, which will affect Y and Z".

Example value: "Describe architecture: how components are connected, how data flows, what are integration points."

code_context - Code context

Purpose: Describes how AI should explain code context that is NOT attached to task, but is relevant.

How to use: Sometimes you don't want to send 50 files, but AI should know they exist. This field allows AI to describe "what's out of view" - external APIs, services, modules that code interacts with.

Example value: "Describe context of code not attached to task: external APIs we use, backend services, modules we integrate with."

← Previous: Extension Back to guide