v1.6.0

Added

• Support for Llama 3.2 model
• Set Llama 3.2 as the default model
• New system prompt to improve quality of generated responses
• Add project name to the first user message in a conversation
• New /model <modelName> command to switch between 'llama3.1' and 'llama3.2'

Full Changelog: https://github.com/fmueller/jarvis/commits/v1.6.0

v1.5.0

Added

• Plain chat command to send a message without code context

Changed

• Always add selected code to the prompt, even if no flag is provided

Full Changelog: https://github.com/fmueller/jarvis/commits/v1.5.0

v1.4.0

Added

• Moving token window to respect the context limit

Changed

• Use LangChain4j as the default LLM framework

Full Changelog: https://github.com/fmueller/jarvis/commits/v1.4.0

v1.3.0

Fixed

• Compatibility with the latest Jetbrains Platform version

Full Changelog: https://github.com/fmueller/jarvis/commits/v1.3.0

v1.2.0

Added

• Streaming updates from the LLM
• Sensible timeouts for the Ollama client

Changed

• Use LLama 3.1 model

Full Changelog: https://github.com/fmueller/jarvis/commits/v1.2.0

v1.1.0

Added

• Support for selected code in prompts

Full Changelog: https://github.com/fmueller/jarvis/commits/v1.1.0

v1.0.0

Added

• Ollama integration to query LLMs
• Chat interface to list messages and enter a new message
• Support for rendering Markdown
• Support for rendering code blocks with syntax highlighting
• Support for slash commands
• Command to start a new conversation
• Help command to list all available commands
• Support for different themes

Full Changelog: https://github.com/fmueller/jarvis/commits/v1.0.0