olimorris/codecompanion.nvim

Features

• Editor-Integrated Chat Interfaces - Opens a chat buffer inside the editor to converse with language models and receive coding assistance.
• Streaming Chat Responses - Opens a turn-based chat buffer within Neovim for conversing with an LLM, with responses streamed back in markdown.
• Inline Code Transformations - Selects code and applies AI-driven edits like refactoring, creation, or modification directly in the buffer.
• System-Level Instructions - Provides system-level instructions to shape LLM responses in chat buffers.
• AI Provider Adapters - Connects to any LLM provider, including self-hosted models, through a flexible adapter system.
• Inline Code Prompts - Sends a prompt that writes or edits code directly into the current Neovim buffer, optionally using visual selection or prompt aliases.
• Editor Context Injections - Dynamically inserts editor state like buffer contents and diagnostics into AI prompts.
• Chat Message Injections - Inserts relevant Neovim context like buffer contents, diagnostics, or viewport into chat messages using a # syntax for dynamic inclusion.
• Editor Buffer Context Providers - Sends the contents of all currently open buffers to the LLM, automatically filtering out excluded buffer types.
• Chat Model Interfaces - Opens a chat buffer inside Neovim to converse with language models and agents for coding assistance.
• LLM Provider Adapters - Overrides the default LLM adapter for a single inline prompt by including the adapter name in the command.
• Natural Language Code Editing - Edits code directly in a buffer by sending a prompt and optional selection to an LLM.
• Editor-Integrated Agents - Connects to agentic tools through the Agent Client Protocol to automate multi-step coding tasks.
• Editor-Integrated Chat Interfaces - Opens a dedicated buffer for back-and-forth conversation with a language model, supporting editor context, slash commands, and tools.
• Autonomous Coding Workflows - Combines a workflow with tools so an LLM can autonomously edit code, run tests, and iterate based on results.
• Editor Buffer Inserters - Inserts the content of any open Neovim buffer into the chat buffer, with support for multiple selections and picker integrations.
• Chat-Integrated Command Execution - Tags a command in the chat buffer so the LLM can execute it on the machine and report results.
• Editor Buffer Sharing - Shares the current buffer's content with the LLM, optionally syncing changes on each turn for iterative feedback.
• Microphone Input Buffering - Shares all open buffers with the LLM, automatically filtering out excluded buftypes and filetypes.
• Selection Interfaces - Shares the current or most recent visual selection with the LLM for targeted code analysis.
• Code Selection Sharing - Sends the current or most recent visual selection to the LLM, useful for asking about specific code.
• Buffer Content Exporters - Sends the content of a specific buffer to the LLM, with parameters controlling how it is synced on each turn.
• AI Code Change Acceptances - Accepts or rejects code modifications proposed by the LLM using configurable keymaps.
• Context Sharing - Sends file contents, buffer diffs, or other data to the LLM as context, with options to sync updates on each turn.
• MCP Server Management - Starts and stops Model Context Protocol servers from within a chat buffer, with the change applying globally.
• MCP Server Connections - Configures external MCP servers to provide tools and context for AI-assisted coding workflows.
• AI Chat Actions - Adds keymaps, slash commands, and tools to the chat interface through an extension's setup function.
• Chat Response Navigations - Moves between LLM responses in the chat buffer using bracket keybindings.
• Runtime Switching - Changes the active LLM adapter and model mid-conversation from within the chat buffer via a selection window.
• Prompt Customization - Defines reusable prompts, editor context, and slash commands tailored to personal workflows.
• Code Explanation - Sends selected code to an LLM to produce a plain-language explanation of how it works.
• Vision Model Inputs - Includes images or screenshots as input in a chat message for vision-capable models.
• Project Context Rules - Reads rule files to guide how the model responds according to project conventions.
• Multimodal Vision Inputs - Sends images as part of a chat message to models that support visual understanding.
• Model Configuration Settings - Edits a YAML block at the top of the chat buffer to adjust adapter hyperparameters before the next LLM turn.
• Prompt Assembly Systems - Constructs a system prompt from a function that receives chat context for dynamic content.
• Prompt Libraries - Runs pre-built prompts for actions like explaining LSP errors or reviewing code.
• Command Palettes - Opens a command palette listing all available plugin actions and built-in prompts for quick access.
• Action Palettes - Opens a searchable palette that lists all available interactions, prompts, and tools for quick selection and execution.
• Code Workflow Executions - Executes a multi-step automated workflow that processes code through a series of actions.
• AI Context Completions - Invokes a completion menu by typing # or @ to insert context variables, tools, or agent-specific commands.
• Git Diff Viewers - Shares the current git diff, including staged and unstaged changes, for code review or commit message generation.
• AI-Context Diff Sharing - Sends the current staged and unstaged git diff to the LLM for code review or commit message generation.
• Chat-Initiated Edit Applications - Tags a file in the chat so the LLM can directly modify code without manual copy-paste.
• Inline Diff Previews - Shows a diff between the original buffer and the LLM's changes, with keymaps to accept or reject the edit.
• Iterative Edit-Test Workflows - Executes a workflow that iteratively edits code and runs tests until they pass.
• Commit Message Generation - Creates a git commit message by analyzing the current buffer's changes through an LLM prompt.
• Chat Session Management - Keeps several independent AI conversations open simultaneously in separate buffers.
• Agent Workflow Executions - Executes a multi-step agent workflow such as code generation or edit-test-repeat.
• Diagnostic Explanations - Sends diagnostic information from active LSP servers in the current buffer to the LLM for troubleshooting.
• AI-Powered Code Fixes - Uses an LLM to analyze and automatically correct errors in selected code.
• AI-Assisted Test Generators - Produces unit tests for the selected code by sending it to an LLM prompt.
• Chat Buffer Visibility Toggles - Opens, closes, or cycles between chat buffers with a command, keeping per-tab chat state when configured.
• Chat Template Customization - Overrides the default system prompt sent with every chat request to tailor AI response style.
• AI Assistants - Delivers an inline assistant experience similar to Copilot Chat.
• IDE and Editor Plugins - Provides an AI-powered coding assistant experience in terminal editors.
• Plugins and Extensions - Neovim plugin for AI-assisted coding with slash commands and agent support.

Historique des stars