Project Roadmap

[mdjastrzebski]

This document describes project roadmap. We plan to release v1.0 at some stage, when some of the features below are implemented. Since this is CLI tool the exact project version should not be as important as in the case of library projects.

Inference providers

• OpenAI
• Antropic
• Perplexity
• Mistral AI
• Gemini
• HuggingFace (?)

Core features

• --model option to select different models
• --provider option to select different providers
• Model parameter: system prompt
• Creativity/precise/default response style option
• ai init command
• /save command for saving current conversation
• Upload text file for context (txt, markdown)

Next steps

• Migration to Ink
• Streaming responses
• Antropic
• Update notifier
• Upload HTML files/URL for querying/summarization

Other backlog

• Personas: create AI assistants for different use cases (e.g. JS, Python, etc)
• Model parameters: temperature, top_p, top_k
• Model parameters: context size
• Upload encoded text files for context (HTML, PDF, doc, etc)
• Upload image file for context
• Read context from URL (this might help)
• Document summarization
• Query available models
• Local indexing database: adding and querying documents
• Save query history (optional)
• Generating images (DALL-E, etc)

[szymonrybczak]

@mdjastrzebski @Q1w1N did you consider streaming response from the external providers? 🤔