codetyper-cli
v0.4.7
Published
CodeTyper AI Agent - Standalone CLI for autonomous code generation
Maintainers
Readme
CodeTyper CLI
An AI-powered terminal coding agent with an interactive TUI. CodeTyper autonomously executes coding tasks using tool calls with granular permission controls and intelligent provider routing.
How It Works
CodeTyper is an autonomous coding agent that runs in your terminal. You describe what you want to build or fix, and CodeTyper:
- Analyzes your request and breaks it into steps
- Executes tools (bash, read, write, edit) to accomplish the task
- Asks permission before modifying files or running commands
- Learns from your project to provide context-aware assistance
Cascading Provider System
CodeTyper uses an intelligent provider routing system:
User Request
|
v
[Detect Task Type] --> code_generation, bug_fix, refactoring, etc.
|
v
[Check Ollama Score] --> Quality score from past interactions
|
v
[Route Decision]
|
+-- High Score (85%+) --> Ollama Only (trusted)
|
+-- Low Score (40%-) --> Copilot Only (needs improvement)
|
+-- Medium Score --> Cascade Mode
|
v
[1. Ollama generates response]
|
v
[2. Copilot audits for issues]
|
v
[3. Update quality scores]
|
v
[Return best response]Over time, CodeTyper learns which provider performs best for different task types.
Installation
Requires Bun runtime (v1.0.0+)
# Install Bun if you don't have it
curl -fsSL https://bun.sh/install | bash
# Install globally via npm
npm install -g codetyper-cli
# Or install from source
git clone https://github.com/CarGDev/codetyper.cli.git
cd codetyper.cli
bun install && bun run build && bun link
# Login to a provider
codetyper login copilot
# Start interactive chat
codetyperFeatures
Interactive TUI
Full-screen terminal interface with real-time streaming responses.

Key bindings:
Enter- Send messageShift+Enter- New line@- Open file picker (works anywhere in input)/- Open command menu (works anywhere in input)Ctrl+M- Toggle interaction modeCtrl+T- Toggle todo panelCtrl+O- Toggle activity panelShift+Up/Down- Scroll log panelCtrl+C(twice) - Exit
Vim Motions
Optional vim-style keyboard navigation for power users. Enable in settings.
Normal Mode:
j/k- Scroll down/upgg/G- Jump to top/bottomCtrl+d/u- Half page scroll/- Search,n/N- Next/prev matchi- Enter insert mode:- Command mode (:qquit,:wsave)
Configuration:
{
"vim": {
"enabled": true,
"startInNormalMode": true
}
}Command Menu
Press / to access all commands organized by category.

Available Commands:
| Category | Command | Description |
| -------- | ------------ | --------------------------------- |
| General | /help | Show available commands |
| General | /clear | Clear conversation history |
| General | /exit | Exit the chat |
| Session | /save | Save current session |
| Session | /context | Show context information |
| Session | /usage | Show token usage statistics |
| Session | /remember | Save a learning about the project |
| Session | /learnings | Show saved learnings |
| Settings | /model | Select AI model |
| Settings | /agent | Select agent |
| Settings | /mode | Switch interaction mode |
| Settings | /provider | Switch LLM provider |
| Settings | /status | Show provider status |
| Settings | /theme | Change color theme |
| Settings | /mcp | Manage MCP servers |
| Account | /whoami | Show logged in account |
| Account | /login | Authenticate with provider |
| Account | /logout | Sign out from provider |
Agent Mode with Diff View
When CodeTyper modifies files, you see a clear diff view of changes.

Interaction Modes:
- Agent - Full access, can modify files
- Ask - Read-only, answers questions
- Code Review - Review PRs and diffs
Permission System
Granular control over what CodeTyper can do. Every file operation requires approval.

Permission Scopes:
[y]Yes, this once[s]Yes, for this session[a]Always allow for this project[g]Always allow globally[n]No, deny this request
Model Selection
Access to multiple AI models through GitHub Copilot.

Available Models:
- GPT-5, GPT-5-mini (Unlimited)
- GPT-5.2-codex, GPT-5.1-codex
- Grok-code-fast-1
- And more...
Theme System
15+ built-in themes to customize your experience.

Available Themes: default, dracula, nord, tokyo-night, gruvbox, monokai, catppuccin, one-dark, solarized-dark, github-dark, rose-pine, kanagawa, ayu-dark, cargdev-cyberpunk, pink-purple
Providers
| Provider | Models | Auth Method | Use Case | | ------------------ | --------------------------- | ------------------- | --------------------------- | | GitHub Copilot | GPT-5, Claude, Gemini | OAuth (device flow) | Cloud-based, high quality | | Ollama | Llama, DeepSeek, Qwen, etc. | Local server | Private, offline, zero-cost |
Cascade Mode
When both providers are available, CodeTyper can use them together:
- Ollama processes the request first (fast, local)
- Copilot audits the response for issues
- Quality scores update based on audit results
- Future requests route based on learned performance
Check provider status with /status:
═══ Provider Status ═══
Current Provider: copilot
Cascade Mode: Enabled
Ollama:
Status: ● Available
Quality Score: 72%
Copilot:
Status: ● AvailableConfiguration
Settings are stored in ~/.config/codetyper/config.json:
{
"provider": "copilot",
"model": "auto",
"theme": "default",
"cascadeEnabled": true,
"maxIterations": 20,
"timeout": 30000
}Project Context
CodeTyper reads project-specific context from:
.github/- GitHub workflows and templates.codetyper/- Project-specific rules and learningsrules.md- Custom instructions for the AI
CLI Usage
# Start interactive TUI
codetyper
# Start with a prompt
codetyper "Create a REST API with Express"
# Continue last session
codetyper --continue
# Resume specific session
codetyper --resume <session-id>
# Use specific provider
codetyper --provider ollama
# Print mode (non-interactive)
codetyper --print "Explain this codebase"Tools
CodeTyper has access to these built-in tools:
| Tool | Description |
| ------------ | ----------------------------------- |
| bash | Execute shell commands |
| read | Read file contents |
| write | Create or overwrite files |
| edit | Find and replace in files |
| glob | Find files by pattern |
| grep | Search file contents |
| lsp | Language Server Protocol operations |
| web_search | Search the web |
| todo-read | Read current todo list |
| todo-write | Update todo list |
MCP Integration
Connect external MCP (Model Context Protocol) servers for extended capabilities:
# Browse and search available servers
/mcp browse # Interactive browser
/mcp search <query> # Search by keyword
/mcp popular # Show popular servers
/mcp categories # List all categories
# Install a server
/mcp install sqlite
/mcp install github
# Manage servers
/mcp status # Show connected servers
/mcp connect # Connect all servers
/mcp disconnect # Disconnect all servers
/mcp tools # List available tools
/mcp add # Add custom serverMCP Browser Features:
- Search by name, description, or tags
- Filter by category (database, web, AI, etc.)
- View server details and required environment variables
- One-click installation and connection
- 15+ verified servers from Anthropic
Extensibility
Hooks System
Lifecycle hooks for intercepting tool execution and session events.
Hook Events:
PreToolUse- Validate/modify before tool executionPostToolUse- Side effects after tool executionSessionStart- At session initializationSessionEnd- At session terminationUserPromptSubmit- When user submits inputStop- When execution stops
Configuration (.codetyper/hooks.json):
{
"hooks": [
{
"event": "PreToolUse",
"script": ".codetyper/hooks/validate.sh",
"timeout": 5000
},
{ "event": "PostToolUse", "script": ".codetyper/hooks/notify.sh" }
]
}Exit Codes:
0- Allow (optionally output{"updatedInput": {...}}to modify args)1- Warn but continue2- Block execution
Plugin System
Extend CodeTyper with custom tools, commands, and hooks.
Plugin Structure:
.codetyper/plugins/{name}/
├── plugin.json # Manifest
├── tools/
│ └── *.ts # Custom tool definitions
├── commands/
│ └── *.md # Slash commands
└── hooks/
└── *.sh # Plugin-specific hooksManifest (plugin.json):
{
"name": "my-plugin",
"version": "1.0.0",
"tools": [{ "name": "custom_tool", "file": "tool.ts" }],
"commands": [{ "name": "mycommand", "file": "cmd.md" }]
}Custom Tool Definition:
import { z } from "zod";
export default {
name: "custom_tool",
description: "Does something",
parameters: z.object({ input: z.string() }),
execute: async (args, ctx) => ({
success: true,
title: "Done",
output: "...",
}),
};Session Forking
Branch and rewind session history for experimentation.
Commands:
| Command | Description |
|---------|-------------|
| /snapshot [name] | Create checkpoint |
| /rewind [n\|name] | Go back to snapshot |
| /fork [name] | Branch current session |
| /forks | List all forks |
| /switch [name] | Switch to fork |
Sessions are stored in .codetyper/sessions/ with automatic commit message suggestions.
Roadmap
v0.4.0 - Agent Completion
The next major release focuses on production-ready autonomous agent execution:
| Feature | Issue | Status | | --------------------------------------- | ----------------------------------------------------------- | ------- | | Plan Approval Gate | #111 | Planned | | Diff Preview Before Write | #112 | Planned | | Execution Control (Pause/Resume/Abort) | #113 | Planned | | Consistent Model Behavior | #114 | Planned | | Quality Gates (TypeScript, Lint, Tests) | #115 | Planned |
Known Issues
- Agent mode executes without waiting for user plan approval
- Different LLM models may behave inconsistently during agent execution
- No diff preview before file modifications
See CHANGELOG for full version history.
Development
# Watch mode
bun run dev
# Type check
bun run typecheck
# Build
bun run build
# Run tests
bun test
# Lint
bun run lintRecent Changes (v0.4.2)
- Session Stats on Exit: Displays detailed session summary when quitting:
- Total API time spent, session duration, code changes (+/-)
- Per-model token usage breakdown with resume command
- Pink Purple Theme: New built-in color theme
- Activity Panel Toggle:
Ctrl+Oto show/hide the activity panel - Image Paste Fix: Fixed race condition where pasted images were silently dropped
- Terminal Exit Fix: Fixed
997;1ngarbage text appearing on exit - @ and / Anywhere: File picker and command menu now work at any cursor position
- Plan Approval Gate: User confirmation before agent executes plans
- Execution Control: Pause, resume, and abort agent execution
- Text Clipboard Copy/Read: Cross-platform clipboard operations with mouse selection
See CHANGELOG for complete version history.
Documentation
- Changelog - Version history and changes
- Contributing - How to contribute
License
MIT - See LICENSE for details.
