Jot CLI - AI Research Assistant

A CLI-based AI agent designed to assist in writing research papers. It gathers context from your project, drafts content, reviews it for academic rigor, and refines it.

Installation

For End Users

In the future the tool should be available to install via package managers like npm or brew. Currently, the only installation method is to download the appropriate binary for your platform from the releases page and add it to your PATH.

For Development

1. Clone the repository:

   git clone https://github.com/Baz00k/jot-cli.git
   cd jot-cli

2. Install dependencies:

   bun install

3. Configure OpenRouter API Key:

   Set your OpenRouter API key using the config command:

   bun run src/index.ts config set-key YOUR_API_KEY

   Get your API key from https://openrouter.ai/

Usage

jot write "Draft a conclusion for the paper summarizing the key findings on LLM reasoning."

Interactive TUI Mode

Launch the interactive Terminal User Interface by running jot without any arguments:

jot

For development, use:

bun dev

Key Features

• Interactive Task Input: Easily draft and refine your prompts.
• Visual Timeline: Track agent actions, tool calls, and review steps in real-time.

Google Antigravity Provider

Jot CLI supports the internal Google Antigravity API as a provider. Antigravity allows you to use SOTA Gemini and Claude models for free.

Authentication

To use this provider, you need to authenticate via OAuth2:

jot antigravity auth

This command will open a browser window for you to sign in with your Google account.

Configuration

You can configure Jot CLI to use Antigravity models for drafting and reviewing:

# Set the writer model
jot config set-writer google/antigravity-gemini-3-flash

# Set the reviewer model
jot config set-reviewer google/antigravity-claude-sonnet-4-5

Model IDs

When using the Antigravity provider, the model ID should include the antigravity- prefix. You can optionally include the google/ prefix as well. The CLI will automatically strip these prefixes when making API requests.

• Example: antigravity-gemini-3-flash or google/antigravity-gemini-3-flash both will use gemini-3-flash using Antigravity API.

Available Antigravity Models

Model	Description
google/antigravity-gemini-3-flash	Gemini 3 Flash (minimal thinking)
google/antigravity-gemini-3-pro-low	Gemini 3 Pro with low thinking
google/antigravity-gemini-3-pro-high	Gemini 3 Pro with high thinking
google/antigravity-claude-sonnet-4-5	Claude Sonnet 4.5 (no thinking)
google/antigravity-claude-sonnet-4-5-thinking	Sonnet with thinking
google/antigravity-claude-opus-4-5-thinking	Opus with thinking

Antigravity provides limited quota for each model. If the quota is exceeded, the model will return an error. You can always use other google account or wait until the quota is reset.

You can check the quota usage by running jot antigravity quota.

Features

• Context Awareness: Reads your project files to understand style and structure.
• Draft-Review-Refine Loop: Ensures high-quality academic output.
• Cross-Platform Configuration: Works seamlessly on Windows, macOS, and Linux.

Recommended Models

This is a list of models that have been tested and found to work well with Jot CLI:

• google/gemini-3-pro-preview / google/antigravity-gemini-3-pro-high
• anthropic/claude-opus-4.5 / google/antigravity-claude-opus-4.5-thinking
• google/gemini-3-flash-preview / google/antigravity-gemini-3-flash
• openai/gpt-5.2-pro
• x-ai/grok-4.1-fast
• moonshotai/kimi-k2-thinking
• z-ai/glm-4.7