MakeShift AI

Conversations

Loading conversations...

Select a conversation

👤

Select a conversation or start a new chat

Conversation Settings

Conversation Details

Title

Notes

Model Settings

AI Model

Loading models... ▼

Technical Level

Non-Technical Developer

Balanced

Response Length

Concise Detailed

Normal

Reasoning Effort

Minimal High

Minimal (~2-3s)

Controls how much the GPT-5 model thinks before responding. Higher effort = deeper reasoning but slower responses.

RAG Settings

Enable RAG (Retrieval-Augmented Generation)

Search Depth

Quick Super Deep

Normal (up to 10 docs/source)

Normal: Retrieves up to 10 documents per source with light pruning of weak results

Show "Context provided to AI" section in responses

When enabled, displays the list of files that were provided as context to the AI

Select Codebases to Search:

Loading codebases...

Source Configuration

Add New Source

Source Type

Project Context (Optional)

Repository URL

Branch

Specific Page IDs or URLs (comma-separated)

Space Keys (comma-separated)

Labels/Tags (comma-separated)

Recursively index child pages

Source Label

File Extensions (blank for all)

Exclude Patterns (comma-separated)

Indexing Paths (optional - leave empty to index entire repository) Enter paths separated by commas or new lines. Examples: src/, docs/api/, components/

Embedding Model

Initializing... 0%

Indexed Sources

Loading sources...

Edit Project Context

Project Context

Describe this project to help the AI provide better responses. Include purpose, technology stack, architecture patterns, domain concepts, or any other relevant context.