Research AI Assistant
Research AI Assistant helps you create a clear, maintainable academic research
project directly from VS Code.
Ask it to organize the open workspace and it creates missing folders for raw,
intermediate, and final data; tables and figures; literature; reusable code;
tests; configuration; and a LaTeX paper. It never replaces an existing file.
Features
- Natural-language messages from the VS Code Command Palette.
- One focused action: format the open workspace as a research project.
- Safe, additive file creation with a preview option.
- A ready-to-edit LaTeX manuscript and BibTeX bibliography.
- API keys stored in VS Code Secret Storage, not workspace settings.
- OpenAI and OpenAI-compatible hosted or local model endpoints.
- No separate Python installation and no extension telemetry.
Get started
Open the folder containing your research project.
Open the Command Palette (Cmd+Shift+P or Ctrl+Shift+P).
Run Research Assistant: Configure API Key.
Run Research Assistant: Ask About This Project.
Enter a message such as:
Organize this workspace as a research project.
Replies appear in the Research Assistant output panel. The extension asks
for confirmation before sending a message because formatting requests can
create files in the workspace.
Generated project structure
project/
├── data/{raw,intermediate,final}/
├── artifacts/{tables,figures}/
├── literature/{pdfs,notes}/
├── literature/references.bib
├── src/
├── scripts/
├── tests/
├── config/
└── paper/
├── main.tex
└── sections/
Model settings
The default model is gpt-4.1-mini. Change Research Assistant: Model in
VS Code Settings to use another model available to your account.
To use an OpenAI-compatible server such as Ollama, LM Studio, vLLM, or another
provider:
- Set Research Assistant: Base URL to its API URL, such as
http://localhost:11434/v1.
- Set Research Assistant: Model to the server's model name.
- Configure an API key only if that server requires one.
The selected model must support Chat Completions tool calling.
Privacy and safety
- The API key is stored using VS Code Secret Storage.
- Messages are sent only to the configured model provider.
- OpenAI Agents SDK tracing is disabled by this extension.
- Formatting is limited to the selected open workspace.
- Existing files are skipped and never overwritten.
Development
npm install
npm run check
npm test
npm run build
Press F5 in this folder to start an Extension Development Host.
To build an installable package:
npm run package
See PUBLISHING.md for the one-time Marketplace setup and
release steps.
License
MIT
