Key takeaways
- Claims 50+ parallel agent capacity — highest among Mac orchestrators
- Native macOS app built on libghostty (Ghostty terminal library) for GPU acceleration
- Open source, free, requires macOS 26 Tahoe (unreleased as of early 2026)
FAQ
What is Supacode?
Supacode is a native macOS app for running 50+ coding agents in parallel with GPU-accelerated terminals via libghostty.
How much does Supacode cost?
Supacode is free and open source. Download via supacode.sh or brew install supacode.
Who competes with Supacode?
Superset, Conductor, Skwad, and Claude Squad offer similar multi-agent orchestration on Mac.
Executive Summary
Supacode is a native macOS app that claims to run 50+ coding agents in parallel, built on libghostty (the terminal library from Ghostty) for GPU-accelerated performance. It's open source, free, and uses a BYOA (Bring Your Own Agents) model.
| Attribute | Value |
|---|---|
| Company | Supabitapp (indie) |
| Founded | 2025 |
| Funding | Bootstrapped |
| GitHub Stars | 249 |
| Headquarters | Unknown |
Product Overview
Supacode positions itself as the "Native terminal coding agents command center" — a fully native macOS app optimized for running many agents in parallel. The technical approach is distinctive: rather than wrapping web technologies like Electron, Supacode is built in Swift using libghostty as the terminal engine.
The app supports any CLI-based coding agent — Claude Code, Codex, OpenCode — running them directly in native terminals without translation layers. Each agent gets its own git worktree for isolation.
Key Capabilities
| Capability | Description |
|---|---|
| Parallel Execution | Run 50+ agents simultaneously |
| Native Performance | libghostty GPU-accelerated terminals |
| BYOA | Works with Claude Code, Codex, OpenCode, any CLI agent |
| Worktree Isolation | Each agent gets isolated git workspace |
| GitHub Integration | Open PRs, see CI checks, fix conflicts |
Product Surfaces
| Surface | Description | Availability |
|---|---|---|
| macOS App | Native Swift desktop app | Beta |
| Homebrew | brew install supacode | Available |
Technical Architecture
Supacode takes a different technical approach than most competitors: it's built as a fully native macOS app using The Composable Architecture (TCA) for state management and libghostty for terminal rendering.
Key Technical Details
| Aspect | Detail |
|---|---|
| Deployment | Local macOS app |
| Terminal Engine | libghostty (GPU-accelerated) |
| Framework | Swift, The Composable Architecture |
| Build System | Xcode + Makefile |
| Open Source | Yes (license in repo) |
Build Process
make build-ghostty-xcframework # Build GhosttyKit from Zig source
make build-app # Build macOS app (Debug)
make run-app # Build and launch
The build requires either downloading prebuilt libghostty binaries or building from source with Zig 0.15+.
Architecture Highlights
- libghostty integration — Uses the same terminal library that powers Ghostty, providing native-quality rendering
- TCA (The Composable Architecture) — Point-Free's architecture for predictable state management
- No Electron — Pure native Swift, avoiding the overhead of web-based desktop frameworks
Mitchell Hashimoto (creator of Ghostty) has highlighted Supacode as one of the notable projects using libghostty.
Strengths
- Maximum Parallel Capacity — 50+ agents is the highest claimed capacity among Mac orchestrators, suitable for aggressive parallel workflows
- True Native Performance — libghostty provides GPU-accelerated rendering that keeps up with heavy agent output; no Electron overhead
- Open Source Transparency — Full code visibility, community contributions welcome; maintainer reviews every PR personally
- Zero Translation Layer — Agents run directly in native terminals; no PTY wrappers or emulation that could introduce bugs
- Modern Swift Architecture — Built on TCA (The Composable Architecture), making the codebase predictable and testable
Cautions
- Requires macOS 26 Tahoe — The unreleased macOS version is required; not usable on current macOS versions as of early 2026
- Alpha/Beta Quality — Limited releases, small user base (249 stars); expect rough edges
- Build Complexity — Requires building libghostty from Zig source or downloading prebuilt binaries
- Limited Documentation — GitHub README covers basics; no comprehensive docs or tutorials
- Single Developer — Appears to be maintained by one person (khoi), creating sustainability questions
Pricing & Licensing
| Tier | Price | Includes |
|---|---|---|
| Open Source | $0 | Full app, open source |
Licensing model: Open source (license in repository)
Hidden costs: Requires AI coding CLI subscriptions (Claude Code, Codex, etc.)
Competitive Positioning
Direct Competitors
| Competitor | Differentiation |
|---|---|
| Superset | Supacode claims 50+ agents vs Superset's 10+; Supacode is native Swift vs Electron |
| Conductor | Supacode is open source and higher capacity; Conductor is simpler but limited to 2 agents |
| Skwad | Supacode focuses on scale; Skwad focuses on MCP-based agent coordination |
When to Choose Supacode Over Alternatives
- Choose Supacode when: You want maximum parallel capacity with native performance on macOS Tahoe
- Choose Superset when: You want mature UX with built-in diff viewer and PR workflow on current macOS
- Choose Conductor when: You want the simplest possible setup with no configuration
Ideal Customer Profile
Best fit:
- Developers with access to macOS 26 Tahoe (beta testers, early adopters)
- Users who prioritize native performance over cross-platform
- Teams running many parallel agents who hit limits on other tools
- Open source contributors interested in Swift/libghostty
Poor fit:
- Anyone on macOS 15 or earlier (tool won't run)
- Developers who prefer stable, well-documented tools
- Teams needing Windows or Linux support
- Users who want turnkey installation
Viability Assessment
| Factor | Assessment |
|---|---|
| Financial Health | Uncertain (indie/bootstrapped) |
| Market Position | Niche (Tahoe-only limits adoption) |
| Innovation Pace | Active (frequent releases) |
| Community/Ecosystem | Growing (249 stars, 18 forks) |
| Long-term Outlook | Conditional (depends on Tahoe GA) |
Supacode's viability is heavily tied to macOS 26 Tahoe's release. Once Tahoe becomes the current macOS version, the user base could grow significantly. The technical foundation (libghostty, TCA) is solid, but the single-maintainer model creates risk.
Bottom Line
Supacode is a promising but early-stage tool for power users on the macOS bleeding edge.
Recommended for: Developers on macOS 26 Tahoe who want maximum parallel agent capacity with native performance.
Not recommended for: Anyone on current macOS versions, or teams needing production-ready stability.
Outlook: If macOS 26 Tahoe ships in late 2026, Supacode could become a leading option for high-throughput agent workflows. The native Swift + libghostty architecture is technically superior to Electron-based alternatives. Success depends on building community momentum before competitors catch up on performance.
Research by Ry Walker Research • methodology