YOUR AI ASSISTANT · FREE & OPEN SOURCE

Agent Nekko

AI help on your computer.
For coding and everyday work.

Get help writing, coding, and handling repeat tasks. Use AI running on your computer or connect an online AI service. You choose what Nekko can access and review its work.

No account needed to use AI on your computer · Works offline after setup · MIT licensed

MEET NEKKO

AN EXAMPLE OF NEKKO HELPING WITH CODE

Add a dark-mode toggle to the settings page
On it, editing SettingsView.tsx and wiring the theme store…
edit_file · SettingsView.tsx · review changes
Context · 1,284 tok
AGENTS.mdguideline
SettingsView.tsxfile
Prefers tabsmemory

FIVE WAYS TO USE NEKKO · VOICE CONTROL IS PLANNED

What you can do with Nekko

A model is the AI you choose. Local means it runs on your computer, rather than an online service. Start with the tools below, and see what we are planning next.

01 · YOUR AI OPTIONS

Use AI on your computer or online

Chat and work in the Nekko app with an AI model installed on your computer, or connect an online service such as Anthropic or OpenAI. You can also use Nekko from supported coding tools.

Connect your coding tools
02 · THE RIGHT TOOL

Choose the right AI for the job

Use a powerful online model for a difficult plan, then a model on your computer or a cheaper service for smaller tasks. Running AI on your hardware avoids per-token API charges. Your hardware, electricity, and any paid services still cost money.

Choose the model for each chat or workflow today. The unreleased rebrand source also lets you choose a provider and model for delegated tasks. Automatically choosing the best model is still a future direction.

03 · REPEAT TASKS

Automate repetitive work

Save a series of steps as a workflow, schedule it, and check the results in one place. Set up recurring checks, called watchers, for work you want Nekko to keep an eye on.

Your computer and the app running the workflow need to stay on for scheduled tasks and checks.

04 · MODEL SETUP

Manage your AI models in one place

Connect Ollama, LM Studio, or vLLM, the software that runs AI models on your computer. Use the model controls each supports, keep an existing setup, and add online AI services when you need them.

See what you can manage
05 · VOICE CONTROL Planned

Control your computer by voice

We are planning AI that understands your instructions and speaks back, all on your computer. Voice input would use speech recognition on your computer too. You would approve which apps and actions it can use, confirm sensitive steps, and have a visible stop control.

Guided setup would choose suitable AI and speech models by checking your processor (CPU), memory (RAM), graphics or unified memory, free storage, operating system, supported AI software, and each model's minimum requirements. Preview the plan, then approve one install click. You would still need to approve required operating-system permission prompts.

Planned, not available today. Initial downloads would need internet or a prepared offline bundle. After setup, supported actions are intended to work offline, with no cloud fallback. Current builds can work with files and terminal commands and show hardware information, but do not yet offer voice, desktop control, or this setup recommender.

See what we are planning

AI ON YOUR COMPUTER

Run AI on your own computer

Install a model runner such as Ollama, LM Studio, or vLLM separately, then connect it to Nekko. A runner is the software that loads and runs your chosen AI. Nekko adds tools for your files and supported model controls; it does not bundle or automatically install the runner today.

  • Find Ollama, LM Studio, and vLLM running on your computer
  • Download, load, and unload Ollama models from Nekko
  • Load and unload LM Studio models when its lms command-line tool is installed on your computer
  • See which models are loaded and check usage. Stop supported model servers; vLLM models stay loaded until their server stops.
  • Also connect online services such as Anthropic, OpenAI, OpenRouter, or another OpenAI-compatible provider

Example setup, not a scan of your machine

Ollama● connectedlocalhost:11434
LM Studio● connectedlocalhost:1234
vLLM○ not runninglocalhost:8000

UNDERSTAND AND REVIEW THE WORK

See what your AI is doing

An AI agent can do more than reply: it can use tools to work on a task. Nekko shows what information it uses, lets you set limits on its tools, and helps you review changes before you keep them.

01

Choose the information it uses

Context means the information sent with your request. Check what is included and leave out what is not needed.

  1. Context Inspector See the files, instructions, and saved notes included in your next request. Toggle or pin them and see how much space they take up, measured in tokens, the small pieces of text AI models process.
  2. Find relevant files Search across the project folders you choose, so Nekko can find useful code and notes without sending every file to the AI.
  3. Guidelines and memory Give Nekko project instructions through AGENTS.md or CLAUDE.md, and keep saved notes for your preferences. Turn individual instructions and notes on or off.
  4. Save the plan Turn the conversation into a project description in SPEC.md, plus a plan and task list that can guide later work.
  5. Improve your request Get suggestions for making your instructions clearer before you send them. This draft check runs on your computer without an online AI service.
02

Give it useful tools

Connect the tools and repeatable steps your work needs, rather than copying everything into a chat.

  1. Split up a task Delegate smaller pieces of work to separate AI sessions, called sub-agents. Each gets its own instructions, with limits on further delegation and a view of the work in progress.
  2. Reusable instructions Skills are saved procedures for work such as research, planning, code review, and testing. See their steps as a workflow, or install more from the skills marketplace.
  3. Connect coding tools MCP is a standard for connecting AI tools. Nekko can use other MCP servers or act as one. Connect Claude Code, Cursor, Codex, or OpenClaw where their CLI or MCP configuration supports it. These are configurable connections, not bundled integrations.
  4. Connect your services Bring relevant information from Linear, Slack, Discord, Gmail, and Drive into your request instead of pasting it by hand. These online connections need a network.
  5. Longer tasks with limits Give Nekko a goal to plan and work through. Set limits on time or steps, and get a request for guidance if it stops making progress.
03

Review and correct changes

AI can make mistakes. Approval settings, file limits, and change review help you stay involved.

  1. Command approvals Choose which terminal commands are allowed, need approval, or are blocked. The app checks for risky operations such as deleting files or changing system permissions.
  2. Limit file access Keep file tools within the folders you selected. For advanced setups, Docker can provide additional isolation for terminal commands.
  3. Review file changes See edits line by line in the Changes panel. Keep or undo individual lines, whole files, or all changes from the app.
  4. Try again from an earlier step If an AI returns an empty reply, Nekko retries and then reports the problem. Use "Reset here" to return to an earlier message and try a different approach.

Meet Nekko: a small outline cat keeping you company while your agent gets to work.

YOUR FILES AND CONNECTIONS

Choose what stays on your computer

ON YOUR COMPUTER

Work without an internet connection

Once your AI model and its runner are installed, supported tasks can work offline. Read, edit, and search your chosen folders without sending requests to an online AI service.

ONLINE SERVICES

Connect online AI when you need it

Online AI models, downloads, and connected services need internet access and may send data to the providers you configure. You choose those connections and are responsible for their costs.

GETTING STARTED

Choose how to use Nekko

Download the desktop app to get started. If you prefer a technical setup, run Nekko in your browser, in Docker, or from a terminal.

01

Desktop app

Download for macOS, Windows, or Linux.

Get the app
02

In your browser

Run it yourself from a source checkout: npm install, then npm run web. Tasks using AI on your computer can work offline after setup.

Setup guide
03

In Docker

For container-based setups, run docker compose up from a source checkout and connect your project folders and model servers.

Compose setup
04

From other tools

The CLI is Nekko's command-line interface. Use the current kotrain command, or kotrain mcp to connect tools that support the MCP standard.

Client setup

The rebrand source adds an agent-nekko executable, but it is not yet in published packages. Use current Kotrain commands and installers until release. The planned npm package is agent-nekko, not yet published.

FUTURE IDEAS · NOT AVAILABLE TODAY

What we are planning next

Alongside planned voice and computer control, we are exploring ways to share work across a team. These ideas have no committed delivery date.

Exploring

Share work across team computers

Use spare team computers to divide up jobs, with shared workflows to organize the steps. This is a future direction, not an available team-computing service.

Exploring

Run project checks on your machines

Explore a local CI and task-runner alternative for repeat jobs such as code tests, with possible cloud runners for extra capacity. This is not a promise of GitHub Actions compatibility.

READY WHEN YOU ARE

Download Agent Nekko

Free, open-source software. Pick the build for your machine. Releases currently carry the Kotrain name; model providers and hardware may have their own costs.

All builds live on GitHub Releases.

Or build from source: git clone · npm install · npm run dev

Current macOS builds are unsigned. Only if you trust the downloaded release, clear quarantine with xattr -cr "/Applications/Kotrain.app". Read the install guidance

Windows builds are unsigned for now (SmartScreen: More info → Run anyway); free code signing by the SignPath Foundation is in progress.