NEKKO LABS Β· OPEN SOURCE Β· MIT

KOTRAIN

Train your models. Run your agents.
Own your machine.

An open-source, local-first AI coding & cowork app. Point it at Ollama, LM Studio, or vLLM in one click, or bring your own cloud keys, and let a model that can actually see and touch your codebases do the work.

No telemetry Β· no account Β· your code never leaves your machine unless you say so

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

01 Β· THE SURFACE

Everything in one calm window

πŸ’¬

Unified chat / cowork / code

No mode switching. Ask a question or hand off a task, same window, same thread.

πŸ”

Context inspector

See exactly what enters the prompt, files, guidelines, memory, with live token counts. Toggle and pin anything.

πŸ›‘

Guardrails by default

Risky commands (rm -rf, force push, curl|sh) prompt before running. Fully configurable: allow, ask, or deny.

πŸ“¦

Sandbox out of the box

File access jailed to your folders. Optional Docker isolation for shell commands.

πŸ—‚

Multi-folder index

Add multiple roots. Kotrain indexes files and symbols and searches across them instantly.

🧠

Memory management

Global and per-project memory, stored as plain markdown you can read and edit.

πŸ”Œ

Connectors

Linear, Slack, Discord, Gmail, and Google Drive, pull real context into the conversation.

πŸ₯‹

A mascot who trains

Nekko peeks in wearing kendo gear: punches, kicks, and bokken swings while the model thinks.

02 Β· THE HEADLINE FEATURE

Local models, finally first-class

Most tools bolt local models on as an afterthought. Kotrain is built around them.

  • Auto-discover a running Ollama, LM Studio, or vLLM on localhost
  • Pull, load, and unload Ollama models from the app
  • Manage servers remotely and watch token usage over time
  • Bring cloud keys too, Anthropic, OpenAI, OpenRouter, any OpenAI-compatible endpoint
Ollama● connectedlocalhost:11434
LM Studio● connectedlocalhost:1234
vLLMβ—‹ not runninglocalhost:8000

03 Β· WHY KOTRAIN

Not a chat box. Not a blind terminal.

πŸ’¬β†’πŸ› 

LM Studio runs models. Kotrain runs with your work.

Local model UIs are a chat box around a model, no awareness of your files. Kotrain reads, edits, searches, and runs inside your real codebases: multi-folder index, inline editing, per-project memory, guardrails. Same one-click local setup, but the model actually does the work.

βŒ¨οΈβ†’πŸ‘

The power of an agentic CLI, with eyes.

Terminal agents are powerful but blind, you can't see what changed without git diff, and editing means leaving the tool. Kotrain gives an IDE-like surface: browse the tree, view files and diffs, edit inline, with every action visible through the Context Inspector and approvals.

04 Β· SAME APP, ANYWHERE

Run it your way

One engine and one UI across every runtime, pick what fits.

πŸ–₯

Desktop

Native app for macOS, Windows, Linux.

● Available
🌐

Self-hosted web

npx kotrain, the same UI in your browser, fully offline.

● Available
🐳

Docker

docker compose up, workspaces as volumes, local models via host network.

● Available
☁️

Kotrain Cloud Paid

Hosted, with an always-on Zero-Data-Retention mode, cloud history + files, and your phone driving your local model via a secure E2E relay.

β—‹ Planned

Kotrain Cloud keeps inference and tools on your machine, the relay is an end-to-end-encrypted pipe to your own paired agent, so using your local model stays private by design.

05 Β· TOUCHDOWN

Download Kotrain

Free and open source. Pick your platform, builds are published on GitHub Releases.

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

macOS says it's "damaged"? It's just unsigned, run xattr -cr "/Applications/Kotrain.app" once, then open. Details