Stop switching.
Build with AI.
The all-in-one native Mac IDE where AI edits can't wreck your project. Replaces Xcode, Android Studio, VS Code, and Cursor — iOS Simulator, Android Emulator, LLDB, Gradle, and Claude Code, Codex, and Gemini CLI all built in. No Electron.
macOS 15+ · Apple Silicon & Intel · xcodebuild & Gradle · iOS Simulator & Android Emulator · LLDB + Kotlin debugger · Claude Code, Codex & Gemini CLI · Zero telemetry
Highlights in v2
Continuously improving,
every week.
Diff Highlighting
Inline diff views colour adds, deletes, and modified regions with per-hunk accent bars. Reviewing agent changes now reads like a proper review, not a wall of mono text.
Go & Rust Syntax
Tree-sitter grammars for Go and Rust join the existing Swift, Kotlin, JS/TS, Python, and Java support. Full highlighting, symbol outline, and codebase index participation.
Resizable Panels
Drag any divider — sidebar, AI chat, terminal, bottom panel — to the width you want. Layouts are remembered per workspace and restored on reopen.
Font Picker
Pick any monospaced font on your system from Settings. Ligatures, weight, and line height are all adjustable; the whole editor, terminal, and diff surface updates live.
Inline Chat Bar
Press the shortcut over any selection to open a focused chat right in the editor gutter — no panel switch, no context loss. Great for quick refactors and targeted questions.
Chinese Site & Localization
Full Simplified Chinese marketing site at /zh and in-app localization for menus, settings, and the AI chat chrome. English and Chinese ship side-by-side.
lingcode CLI
Drive LingCode from any shell. Interactive REPL with Claude's full Agent SDK — or chat against OpenAI, Groq, Ollama, and any OpenAI-compatible endpoint. Docs →
The Xcode side
A real build system—not just a text editor.
LingCode wraps the full xcodebuild toolchain. Build, run, debug, sign, and ship — all without opening Xcode.
- Build & run — xcodebuild with auto-detected schemes, destinations, and sanitizers. One click.
- iOS Simulator — boot, install, launch, and stream live logs. Full simctl integration.
- LLDB debugger — breakpoints, stack frames, variables, watch expressions, REPL. DAP protocol.
- Code signing — pick your team, toggle automatic/manual signing, manage entitlements & capabilities.
- Test & coverage — run tests, parse results, show green/red coverage in the gutter.
- Archive & ship — archive for Development, Ad Hoc, or App Store. Export and upload to TestFlight.
- SPM packages — add, remove, update Swift packages visually.
- Scheme editor — edit environment variables, launch arguments, build targets.
- Native Mac — AppKit & SwiftUI. Handoff, widgets, Shortcuts, Finder integration, AppleScript.
iPad app available separately for chat & remote workflows—the Mac build is the full IDE.
The Android side
Full Android toolchain, same IDE.
Gradle builds, AVD emulators, adb, logcat, and a Kotlin debugger — in the same native Mac app. No Android Studio required.
- Gradle build — auto-detected modules, variants, and product flavors. Build, install, and run from the toolbar.
- Android Emulator — list, boot, and install on AVDs via
adb; livelogcatstream in the console with tag filtering. - Kotlin debugger — JDWP-backed DAP session with breakpoints, stack frames, variables, and expression eval.
- APK & AAB — debug and release builds; signed AAB output ready for Play Console.
- Manifest-aware — parses
AndroidManifest.xmlfor permissions, activities, and launch intents. - Google Play deploy — service-account upload with preflight checks. Setup guide.
- Device list — physical devices over USB or Wi-Fi debugging; same destination picker as iOS.
Swift, Kotlin, or both — one IDE, one AI, one codebase index.
See it in action
Workflows, not a feature laundry list.
Chat with an agent that actually sees what you see
Build errors, sim logs, debugger state, active file, git, credentials — the agent reads it all, live. No shell-out, no stale stdout. Say "ship this to TestFlight" in chat, watch it go.
Short overview
Native AI coding in under 2 minutes.
Multiple Claude Code Sessions
Run several Claude Code tabs at once. Switch between them without losing your session.
Magic Deploy
Paste a token. LingCode figures out the rest.
Magic Install
Detects npm, pip, Cargo, CocoaPods, and 20+ more. One click.
Magic Push
Stage, AI commit message, push. All from the git panel.
Trust pillar
Your project survives the AI.
AI edits can silently break an Xcode project, corrupt an AndroidManifest.xml, or flatten a Gradle config. LingCode assumes they will, and makes every change reversible before you have to notice.
Pre-edit snapshots
Before the agent writes a byte, copies of project.pbxproj, Info.plist, build.gradle, AndroidManifest.xml, and *.entitlements land in ~/Library/Application Support/LingCode/WorkProtection/. One-click recovery panel with previews — no git required.
Per-hunk accept & revert
Every AI edit renders as a syntax-highlighted diff card on the file. Keep or revert each hunk independently — never all-or-nothing. Review a 40-file refactor without handing over control.
Semantic time-travel undo
A separate multi-file undo stack for AI operations, independent of the editor's native undo. Rewind a cross-file refactor in one step, even after you've kept typing.
Plus worktree isolation for risky experiments, read-before-write enforcement, and a five-layer safety gate before any agent action. See the full trust model →
Xcode power + Cursor intelligence
Everything you need to build,
edit with AI, and ship.
Complete feature list (Mac & iPad)
Agent mode
60+ native Swift tools — builds, runs, debugs, reads your repo, and edits files directly. You approve each step. 3-tier token escalation, auto context compaction, and structured memory across sessions.
Rules (.mdc)
WORKSPACE.md, .cursorrules, .cursor/rules—drop in your Cursor rules; no migration.
Magic workflows
Install — detect npm, pip, Cargo, CocoaPods, and more. Git — stage, AI commit message, push from the git panel. Deploy — guided deploy to common hosts from tokens or SSH. Three demos above show each path.
Skills
Slash commands & workflow scripts: /review, /commit, /security, /mode, /help, plus your own custom ones. Drop-in Cursor compatibility — LingCode reads .cursor/skills, .agents/skills, and ~/.cursor/skills SKILL.md files directly. Type / and go. See the full list
Extensions & MCP
VS Code-style marketplace for LSPs and tools. Bring your own MCP servers (stdio, HTTP, SSE, or WebSocket) — their tools and resources show up alongside the 60+ native ones.
Multiple Claude Code sessions
Parallel Claude Code tabs—each session is independent. Workflow guide
Worktree isolation
Send the agent into an isolated git worktree for risky experiments. Merge, discard, or keep the branch when it's done — your main tree stays untouched. See the full trust model above.
Per-hunk accept & revert
Every AI edit renders as a diff banner on the file. Keep or revert each hunk independently — not all-or-nothing. Green highlighting on live edits, syntax-aware side-by-side view.
Resizable everything
Drag sidebar, AI panel, bottom panel, and Git panel dividers to any size. Layout persists across launches — LingCode remembers your workspace.
AI chat & inline edit
Four chat modes — Agent (autonomous, approval-gated), Plan (design before editing), Ask (Q&A, no edits), Debug (error-focused). Project-aware context, semantic search, diff before apply. OpenAI, Anthropic, DeepSeek, Gemini, Qwen, Kimi — your keys.
LSP IntelliSense + FIM ghost completions
Real language servers (SourceKit, Kotlin, TS, Python, Go, Rust) for prefix completions, diagnostics, go-to-definition. Plus Copilot-style AI ghost text with Fill-in-Middle prompting — the completion sees both what you've typed and what follows the cursor. Against your cloud or local model. Tab to accept, Escape to dismiss.
HTTP proxy inspector
Localhost forward proxy for debugging plaintext HTTP from your app — no root CA install. Capture method, URL, headers, body, and status live. The agent can read recent transactions when diagnosing flaky requests.
Light & dark themes
Full light mode across every surface—editor, terminals, AI chat, diff views, approval dialogs. Terminals get a dedicated light ANSI palette. Follows system appearance or pick manually.
Session persistence
AI tabs and their conversations survive quit & relaunch. Tab state is flushed on Cmd+Q; history is restored on reopen. Pick up exactly where you left off.
Why choose?
Xcode + Android Studio +
VS Code + Cursor.
One app.
Stop juggling four apps. LingCode does it all — with Claude Code, Codex, and Gemini CLI built in.
| Feature | LingCode | Xcode | Android Studio | VS Code | Cursor |
|---|---|---|---|---|---|
| Apple build & ship | |||||
| xcodebuild integration | ✓ | ✓ | ✗ | ✗ | ✗ |
| iOS Simulator | ✓ | ✓ | ✗ | ✗ | ✗ |
| LLDB debugger | ✓ | ✓ | ✗ | ✗ | ✗ |
| Code signing & teams | ✓ | ✓ | ✗ | ✗ | ✗ |
| Archive & App Store export | ✓ | ✓ | ✗ | ✗ | ✗ |
| SPM package manager | ✓ | ✓ | ✗ | ✗ | ✗ |
| Scheme editor | ✓ | ✓ | ✗ | ✗ | ✗ |
| Code coverage gutter | ✓ | ✓ | ✗ | ✗ | ✗ |
| Build settings editor | ✓ | ✓ | ✗ | ✗ | ✗ |
| Entitlements & capabilities | ✓ | ✓ | ✗ | ✗ | ✗ |
| Android build & ship | |||||
| Gradle build & install | ✓ | ✗ | ✓ | ✗ | ✗ |
| Android Emulator (AVD) | ✓ | ✗ | ✓ | ✗ | ✗ |
| Kotlin / JDWP debugger | ✓ | ✗ | ✓ | ✗ | ✗ |
adb & logcat streaming | ✓ | ✗ | ✓ | ✗ | ✗ |
| Signed AAB for Play Console | ✓ | ✗ | ✓ | ✗ | ✗ |
| APK Analyzer | ✓ | ✗ | ✓ | ✗ | ✗ |
| AI & Agents | |||||
| AI inline editing | ✓ | ✗ | via Gemini | via Copilot | ✓ |
| AI agent mode | ✓ | ✗ | ✗ | Copilot beta | ✓ |
| Claude Code built in | ✓ | ✗ | ✗ | ✗ | ✗ |
| Codex built in | ✓ | ✗ | ✗ | ✗ | ✗ |
| Gemini CLI built in | ✓ | ✗ | ✗ | ✗ | ✗ |
| AI chat with codebase context | ✓ | ✗ | via Gemini | via Copilot | ✓ |
| Multiple AI sessions | ✓ | ✗ | ✗ | ✗ | ✗ |
| Agent drives whole IDE (native tools) | ✓ | ✗ | ✗ | ✗ | ✗ |
| Structured memory (per project, 7 categories) | ✓ | ✗ | ✗ | ✗ | ✗ |
| Cursor-compatible SKILL.md loader | ✓ | ✗ | ✗ | ✗ | ✓ |
| Remote agent bridge (SSE) | ✓ | ✗ | ✗ | ✗ | ✗ |
| Pre/post agent hooks | ✓ | ✗ | ✗ | ✗ | ✗ |
| 3-tier context auto-compaction | ✓ | ✗ | ✗ | ✗ | ✗ |
| System-prompt cache (dynamic boundary) | ✓ | ✗ | ✗ | ✗ | ✗ |
| Defense-in-depth safety (5 layers) | ✓ | ✗ | ✗ | ✗ | ✗ |
| Cross-file autocomplete context | ✓ | ✗ | via plugins | via LSP | ✓ |
| Editor & Platform | |||||
| Native Mac app | ✓ | ✓ | JVM | Electron | Electron |
| Extensions marketplace | ✓ | ✗ | IntelliJ | ✓ | ✓ |
| Multi-language LSP | ✓ | Swift only | JVM-focused | ✓ | ✓ |
| Handoff, widgets, Shortcuts | ✓ | ✗ | ✗ | ✗ | ✗ |
| Zero telemetry | ✓ | ✓ | ✗ | ✗ | ✗ |
| Magic Deploy + Push (token-based, guided) | ✓ | ✗ | ✗ | ✗ | ✗ |
| Checkpoints / timeline rollback | ✓ | ✗ | ✗ | ✗ | ✓ |
| iCloud preferences sync | ✓ | ✗ | ✗ | ✗ | ✗ |
| Startup time | < 1 s | ~3 s | ~10 s | ~2 s | ~4 s |
| Price | Free | Free | Free | Free | $20/mo+ |
Pricing
Your keys. Your compute.
No subscription required.
Everything you need. Bring your own API keys from Anthropic, OpenAI, DeepSeek, Gemini, or any compatible provider.
- ✓ Full agent mode — 60+ native Swift tools
- ✓ LSP IntelliSense (Swift, Python, TS, Go, Rust…)
- ✓ Multiple Claude Code sessions
- ✓ MCP server support + extensions marketplace
- ✓ Agent hooks + 5-layer safety guard
- ✓ Structured memory + auto context compaction
- ✓ Remote agent bridge
- ✓ Magic Deploy, Install & Push
- ✓ Checkpoints, session history & conversation persistence
- ✓ Vim-style modal editing (optional)
- ✓ DAP debugger — breakpoints, step, variables, watches
- ✓ Handoff, widgets, Shortcuts
- ✓ iCloud preference sync
- ✓ Zero telemetry
Managed API credits, team seat management, priority support, and advanced team hooks — for professional teams.
- ✓ Everything in Free
- ✓ Managed API credits (no own key needed)
- ✓ Team seat dashboard
- ✓ Shared project hooks & policies
- ✓ Priority support SLA
- ✓ SSO / SAML
You own 100% of your code, prompts, and outputs. We never train on your data.
Safety by design
Defense-in-depth.
Not an afterthought.
Every agent action passes through five independent gates before execution. No single misconfiguration can let the model run destructive commands.
Read the security model →Tool inputs are validated against JSON schema before reaching any executor. Malformed inputs are rejected immediately.
User-defined shell commands or AI prompts run before every tool call. Teams can enforce custom policies without touching app code.
Allow / always-allow / deny per tool. Destructive actions—kill, delete, team ops—always require explicit approval regardless of settings.
Pipe-to-shell fetches (curl | bash), fork bombs, and obfuscated shell tokens are blocked at the kernel safety layer.
File operations are constrained to the open project root. The agent cannot read ~/.ssh, ~/.aws, or any path outside your workspace.
Why developers switch
Built for Apple silicon
and Apple workflows.
"It actually uses native APIs and the rest of the OS—not another Chromium box with a Mac icon."
iOS Engineer"Magic Deploy: paste token, hit deploy. I used to spend an hour on Vercel config. Now it's 30 seconds."
Full-stack Developer"My API key stays on my machine. No subscription, no telemetry, no 'we're training on your code.'"
Security Engineer"The agent hooks let me auto-approve safe commands and block risky ones without clicking through dialogs all day."
Staff Engineer"Cross-file autocomplete finally understands the types I defined three files over. Feels genuinely intelligent."
Swift Developer"Starts in under a second, uses 5× less RAM than Cursor, and ships with 60+ native tools — including build, run, and debug. My MacBook Air fans never spin up."
Indie Hacker