Agentic coding on your own models

The agentic dev environment
for local LLMs

Describe a task in plain language. Ascora ADE reads your code, plans, edits files, runs commands, and shows a diff before applying — powered by the models you run yourself. Private, offline-capable, and entirely in your control.

Works with LM Studio · Codex · Claude Code · GLM — bring your own keys, or run fully local.

Ascora ADE — interactive demo
Loading the interactive ADE…

One agent, every backend — switch models per task

LM StudioLocal · OpenAI-compatible
CodexOpenAI gpt-5.x
Claude CodeOpus · Sonnet · Haiku
GLM / ZCodeZhipu & friends
Everything an IDE should be

A full development environment, with an agent inside it

Agentic editing

It edits the code — you stay in control

The agent works in a loop: read, plan, edit, run, verify. Every file change is shown as a diff you can approve or reject. Trust it more? Switch to Auto-apply and let it run uninterrupted.

  • Side-by-side diffs before anything touches disk
  • “Ask before changes” or “Auto-apply” per task
  • A running tally of every file the task touched
Edit filesrc/theme.ts+12-3needs review
export function useTheme() {
- const [t] = useState('light')
+ const [t, setT] = useState(stored ?? 'system')
+ useEffect(() => persist(t), [t])
}
Approve Reject
Bring your own model

Private by default, on the models you choose

Ascora ADE speaks to local LLMs through LM Studio’s OpenAI-compatible API, and to Codex, Claude Code, and GLM/ZCode. Pick a backend per project — your code never has to leave your machine.

  • Local-first: runs fully offline against LM Studio
  • Switch provider and model from the composer
  • Per-workspace model memory
Agent backend
LM Studioqwen2.5-coder-7b
Codexgpt-5.4
Claudesonnet
GLM · ZCodeglm-4.6
Connected · runs on your machine
Thinks out loud

See the reasoning, the tools, the result

Follow the agent’s thought process, watch each tool call stream in — list a directory, read a file, search, run a command — and read the final summary. No black box.

  • Collapsible reasoning blocks
  • Live token + time counter while it works
  • Tool cards with output, diffs, and exit codes
Add a dark-mode toggle to the settings page
Thought process3.2k tokens · 4s
Read filesettings.tsxdone
Run commandnpm testdone
Assistant — Done. I added a ThemeToggle and wired it to persisted settings. Tests pass.

Monaco editor

The same editor that powers VS Code, bundled and fully offline.

Integrated terminal

Run builds and tests without leaving the workspace.

Git, with AI commits

Stage, diff, branch, and push — or let the agent write the commit message.

SSH terminals

Drive remote hosts from the same agent loop over SSH.

Live preview

Built-in live server renders your HTML as you change it.

Usage analytics

Track tokens, models, and tasks across every workspace.

Your code, your machine

Nothing is uploaded unless you choose a hosted backend.

Dockable panels

A VS Code-style layout you can split, stack, and re-dock.

From prompt to pull request

Four steps, no context switching

01

Open a project

Point Ascora ADE at any folder. It indexes the tree and connects to your chosen model.

02

Describe a task

Ask in plain language — “add a dark-mode toggle”, “fix the failing test”, “refactor this”.

03

Review the diff

The agent plans and proposes edits. Inspect each change as a clear, reviewable diff.

04

Apply with confidence

Approve to write to disk, or let Auto-apply run the whole task end to end.

Start building with your own models

Free to download. Install Ascora ADE, point it at a project, and ship your first AI-assisted change in minutes.

view the changelog
Questions

Frequently asked

Not unless you choose a hosted backend. With LM Studio, everything — the model and your code — runs locally and works fully offline. Codex, Claude Code, and GLM call their respective services only when you select them.

Any OpenAI-compatible local model through LM Studio, plus the Codex (gpt-5.x), Claude Code (Opus / Sonnet / Haiku), and GLM/ZCode CLIs. You can switch the backend and model per task from the composer, and each workspace remembers its own choice.

Ascora ADE is built around local-first, agentic workflows: it’s designed to run against models you host yourself, with a transparent read-plan-edit-run loop and explicit diff approval. You own the stack end to end.

Only if you let it. The default “Ask before changes” mode shows a diff and waits for your approval on every write and command. Switch to “Auto-apply” when you want it to run uninterrupted.

Ascora ADE is a desktop app built on Electron, so it runs on Windows, macOS, and Linux. The interactive demo on this page runs the real interface right in your browser with mock data.

No — Ascora ADE is proprietary software developed by AS CORE AI. It’s free to download and use, but the source isn’t publicly licensed.