Code faster without giving up control.

NanoAgent runs on your machine. It understands your repository, plans changes, edits files, runs validation, and reviews diffs. Quietly capable, incredibly precise.

$ curl -fsSL https://raw.githubusercontent.com/getnanoai/NanoAgent/master/scripts/install.sh | bash
PS> irm https://raw.githubusercontent.com/getnanoai/NanoAgent/master/scripts/install.ps1 | iex
$ npm install -g nanoai-cli
$ pnpm add -g nanoai-cli
$ bun add -g nanoai-cli
$ ext install rizwan3d.nanoagent

Docked tool window drives the local nanoai CLI over ACP.

Extensions ▸ Manage Extensions ▸ search “NanoAgent”

Docked tool window drives the local nanoai CLI over ACP.

Windows · macOS · Linux (x64 / arm64)

Verify with SHA256SUMS + build attestation.

nanoai
NanoAgent running in a terminal

Works Where You Work

  • VS Code
  • Visual Studio
  • Terminal
  • Desktop
  • CI/CD Workflows

Why NanoAgent?

Because an AI coding assistant should be powerful, transparent, and respect your workspace.

Precision engineered for complex codebases.

NanoAgent isn't a chatbot. It's a structured workflow engine that operates autonomously but transparently.

Repository Understanding

NanoAgent builds a semantic graph of your entire codebase locally. It knows where things are and how they connect.

Strategic Planning

Before writing a single line of code, the agent formulates a step-by-step plan for your approval.

Precision Editing

Applies targeted modifications to files. No messy copy-pasting, just clean, structural diffs.

Automated Validation

Runs your tests and linters in the background, automatically fixing errors until the build goes green.

Human Review

You remain in the driver's seat. Review the proposed diffs before they are committed.

How NanoAgent compares.

A practical look at what makes NanoAgent different from the tools you already know.

Feature NanoAgent Codex Claude Code OpenCode Kilo Code GitHub Copilot Aider
Local-first
Repo memory
VS Code extension
Visual Studio
ACP-compatible editors
CI code reviews
15+ AI providers
ACP protocol
Sandboxed execution
Open source
LSP
Direct shell & background terminals
Git sidebar
Subagents & orchestration
Local codebase indexing
Memory, audit & hooks
NuGet library release

Yes · Partial · No

Comparison is based on publicly documented capabilities as of mid-2026. Features change quickly; verify details on each product's official site before making a decision.

Provider choice

Use the model that fits your budget & policy

From subscription sign-in to API-key providers to fully local models — NanoAgent adapts to what you already pay for.

Your code. Your machine. Zero compromise.

In a world of cloud-hosted AI, NanoAgent takes a different path. Your proprietary source code never leaves your workstation unless you push it.

  • Total Privacy

    Operates entirely within your local environment or secure VPC.

  • Zero Latency

    Local models mean instant responses and no rate limits.

  • Local Indexing

    Vector embeddings are stored locally, avoiding third-party data collection.

One gateway, every model.

NanoAgent Gateway is the control plane for enterprise AI. Put a single OpenAI-compatible endpoint in front of your traffic and give security, finance, and platform teams the visibility and control they need.

Security

Track how teams use models and watch for anomalies as adoption grows.

Finance

Attribute usage by team, app, project, and user so budgets stop being guesswork.

Platform

Keep clients pointed at one endpoint while your team manages changes behind the scenes.

import { streamText } from 'ai'import { createOpenAI } from '@ai-sdk/openai' const nano = createOpenAI({  baseURL: 'https://app.getnanoai.com/v1',  apiKey: process.env.NANOAGENT_API_KEY,}) const result = streamText({  model: nano.chat('anthropic/claude-opus-4-8'),  prompt: 'Why is the sky blue?',})
import osfrom openai import OpenAI client = OpenAI(    base_url="https://app.getnanoai.com/v1",    api_key=os.environ["NANOAGENT_API_KEY"],) stream = client.chat.completions.create(    model="anthropic/claude-opus-4-8",    messages=[{"role": "user", "content": "Why is the sky blue?"}],    stream=True,)
# point any OpenAI-compatible client at the Gatewaycurl https://app.getnanoai.com/v1/chat/completions \  -H "Authorization: Bearer $NANOAGENT_API_KEY" \  -H "Content-Type: application/json" \  -d '{    "model": "anthropic/claude-opus-4-8",    "messages": [{"role": "user", "content": "Why is the sky blue?"}],    "stream": true  }'
NanoForge · Full-stack app builder

Describe what you want. NanoForge builds it.

NanoForge's agent scaffolds, builds, and live-previews a real React + Express + PostgreSQL app in an isolated container, backed by its own GitHub repo and database.

Full-stack generation

React frontend, Express API, and PostgreSQL schema generated together from one description.

Isolated preview environment

Each app runs in its own container with a live URL and hot reload before it touches your machine.

Review every diff

Changes are committed to a dedicated GitHub repo so you can inspect, approve, or edit before merging.

NanoForge full-stack app builder preview

NuGet library

The core NanoAgent package also ships on NuGet. Tagged releases publish the NanoAgent library to NuGet.org alongside the desktop and CLI assets. The end-user command-line experience is still distributed through the release installers.

dotnet add package NanoAgent

Proudly Open Source.

Released under the permissive Apache-2.0 License. Contribute, fork, and build upon NanoAgent without restrictions.