nanocode-cli 0.3.16

A lightweight terminal-based AI coding assistant

nanocode

A lightweight terminal-based AI coding assistant.

nanocode is used to help building itself, including features such as @file path completion.

Screenshots

Features

• Constrained Output: Force model replies into auditable action frames.
• Verified Edits: Reject stale range edits before they touch files.
• Autonomous Loop: Chain reading, editing, running, and verification.
• Live Telemetry: Stream tool intent, token use, and status.

Install

uv tool install nanocode-cli

Upgrade an existing install:

uv tool upgrade nanocode-cli

For local development:

uv sync --extra dev
uv run nanocode

Usage

Start nanocode:

nanocode

Show available commands:

/help

Ask a source-aware question about nanocode itself:

/help how does compact work?

CLI arguments:

• --yolo: Skip tool execution confirmations.
• --debug: Write request prompts to .nanocode/debug.
• --config <path>: Path to config file (default: ~/.nanocode/config.toml).
• --init-config: Create a default config file.
• -v, --version: Show program version.

Safety

nanocode does NOT provide sandbox protection. It can run shell commands and edit files in the environment where you start it.

If you do not fully trust the model, tools, prompts, or workspace, run nanocode inside your own sandbox, container, VM, or other isolated environment.

USE AT YOUR OWN RISK.

Tools

• File: Read, LineCount, ListDir, Search.
• Edit: Edit, ReplaceRange, ApplyPatch.
• Shell: Bash, Git.
• Memory: Recall reads stored tool results by key.

Commands

• Info: /help [question], /status, /rules, /knowledge [update], /compact.
• Config: /config, /set <key> <value>, /model [model_name], /provider [name], /yolo.
• Maintenance: /clean-logs.
• Exit: /exit, /quit.

Configuration

Run nanocode --init-config to create ~/.nanocode/config.toml.

• Provider config: [provider] active = "<name>" plus [provider.<name>] url, key, model, and model options.
• Runtime config: [paths] and [runtime].

Status

• Status bar: active model, reasoning, conversation context, current-turn tool calls, tokens, elapsed time, and active model-call time.
• /status: active provider, model state, runtime state, conversation/tool counters, per-model calls/tokens, task, goal, and verification.