Skip to content

replit.md

Latest commit

 

History

History
91 lines (68 loc) · 2.92 KB

replit.md

File metadata and controls

91 lines (68 loc) · 2.92 KB

Emigo - AI Assistant for Emacs

Project Overview

Emigo is an intelligent, agentic Emacs-native AI assistant that understands and interacts with your codebase. It's designed to run as a Python backend that communicates with Emacs via EPC (Emacs Process Communication).

Key Features:

  • Agentic tool use - interacts with your environment based on LLM reasoning
  • Native Emacs integration
  • Flexible LLM support via LiteLLM (OpenRouter, Deepseek, etc.)
  • Context-aware interactions with chat history and project context

Project Type

This is NOT a web application - it's an Emacs plugin with a Python backend. The Python code (emigo.py) runs as an EPC server that Emacs connects to for AI-powered coding assistance.

Architecture

Main Components:

  1. emigo.el - Emacs Lisp frontend
  2. emigo.py - Python EPC orchestrator (main backend)
  3. llm_worker.py - LLM interaction subprocess
  4. session.py - Session management
  5. tools.py - Tool implementations (read_file, write_file, execute_command, etc.)
  6. repomapper.py - Repository mapping for context

How It Works:

  1. Emacs starts the Python backend (emigo.py) as a subprocess
  2. Python starts an EPC server on a random port
  3. Emacs connects to the EPC server
  4. User interacts via Emacs buffers
  5. LLM worker handles AI interactions and tool execution

Replit Setup Status

✅ Python 3.11 installed ✅ All Python dependencies installed (litellm, epc, networkx, etc.) ✅ Project is ready to use

Usage in Emacs

This project is designed to be used from within Emacs:

  1. Install in Emacs using straight.el:
(use-package emigo
  :straight (:host github :repo "MatthewZMD/emigo" :files (:defaults "*.py" "*.el"))
  :config
  (emigo-enable)
  :custom
  (emigo-model "openrouter/deepseek/deepseek-chat-v3-0324")
  (emigo-base-url "https://openrouter.ai/api/v1")
  (emigo-api-key (getenv "OPENROUTER_API_KEY")))
  1. Basic Usage:
  • M-x emigo - Start Emigo in your project
  • Type prompts and press C-c C-c to send
  • Use @filename to add files to context
  • C-c f - Add files interactively
  • C-c l - List files in context
  • C-c H - Clear chat history

Testing on Replit

Since this is an Emacs plugin and Replit doesn't have Emacs pre-installed, you can:

  1. Validate Python Setup:

    python3 --version  # Check Python
pip list | grep -E "(litellm|epc)"  # Verify dependencies

  2. Check Code Structure:

    python3 -m py_compile emigo.py  # Syntax check

  3. Install Emacs (optional): If you want to actually run Emigo in Replit, you would need to install Emacs first.

Development Notes

  • Written in Python 3.x (tested with 3.11)
  • Uses EPC for Emacs-Python communication
  • LLM interactions via LiteLLM (supports multiple providers)
  • Active development - expect breaking changes

Recent Setup

  • 2025-10-13: Imported to Replit, installed Python 3.11 and all dependencies