Navamai enhance your craft with ai

NavamAI aims to transform your terminal into a configurable, interactive, and unobtrusive personal AI application. It invokes multiple large language models (LLMs) through simple commands, helping you boost productivity and workflow efficiency.

Main Capabilities

• Multi-model support: Integrates 15 LLMs and 7 providers, including Groq with Llama 3.1, Sonnet 3.5, GPT-4o, Mistral Nemo, Gemini 1.5 Flash, and more, allowing users to freely choose based on speed, capability, context length, or local operation needs.
• Rich commands: Provides 15 commands to customize personal AI workflows, among which navamai, ask, image, and refer are the four top-level commands. Such as ask "create a table of planets" can obtain streaming responses, Markdown tables, and code block highlighting directly in the terminal.
• Markdown workflow automation: Compatible with Markdown content, it can work with tools such as Obsidian, VS Code, Material for MkDocs, and Jekyll to design custom workflows.
• Contextual application generation: Through the ask command, select an application creation prompt template to generate, install, and run a fully functional web application within seconds, without code setup or dependency installation. The generated application is presented as a Markdown blog, including code blocks, file sections, installation and run scripts, and can be viewed and edited in tools such as Obsidian. Use the refer iterate-code command to inline-prompt modifications and iterate, while maintaining the code blog version. Example: with just two commands and in less than a minute, create a stock analysis dashboard with real-time market data.

How to Use

1. Install NavamAI in the folder where you want to initialize NavamAI (such as an Obsidian vault, VS Code project folder, or empty folder).
2. Interact with AI through simple commands (such as ask) to get a chat-style UI within the terminal, supporting streaming responses and Markdown formatting.
3. Switch models, providers, prompts, model parameters, folders, and document types by modifying configuration files or using easy-to-remember commands.
4. All settings have sensible defaults and work out of the box, while also supporting full customization and extension.

Applicable Scenarios

• Scenarios requiring quick search, research, Q&A, or generation of saveable artifacts.
• Automating Markdown writing and documentation workflows.
• Quickly creating contextual web applications (such as a stock analysis dashboard) and iterating.
• Using local models (such as Mistral Nemo) in offline environments.
• Wanting to integrate AI into development, writing, or note-taking processes without leaving the terminal.

Limitations

• The specific list of supported models and providers may change with configuration updates, and some models require corresponding API keys or local setup.
• The quality and context size of contextual application generation depend on the selected model.
• Official information on pricing, platform compatibility details, or performance benchmarks is not yet available.