AeroLaunch transforms your terminal into a multi-dimensional cockpit for AI-powered software development. Unlike traditional single-session tools, this platform lets you run multiple Claude and GPT agents simultaneously across different projects, while maintaining a bird's-eye view of every conversation, code change, and deployment.
Imagine having a dozen expert developers, each specialized in different domains, all working simultaneously on your projects — that's AeroLaunch. This is not another chatbot wrapper. This is a mission control center for AI coding sessions where each "bubble" represents an independent agent with its own context, memory, and toolset.
Built on the official Claude Agent SDK and engineered for parallel processing, AeroLaunch breaks the single-threaded bottleneck of standard AI coding assistants. Whether you're debugging a microservice in bubble-1, generating unit tests in bubble-2, or reviewing a pull request in bubble-3, every operation runs concurrently without context bleeding.
graph TB
subgraph "AeroLaunch Engine"
A[User Command Center] --> B[Session Orchestrator]
B --> C[Bubble-1: Frontend Agent]
B --> D[Bubble-2: Backend Agent]
B --> E[Bubble-3: Database Agent]
B --> F[Bubble-N: Custom Agent]
end
subgraph "Parallel Execution Layer"
C --> G[Claude API]
D --> H[OpenAI API]
E --> I[Claude API]
F --> J[Hybrid API]
end
subgraph "Integrated Tools"
G --> K[Terminal Emulator]
H --> L[Browser Bubble]
I --> M[Database Console]
J --> N[Code Review Panel]
end
K --> O[Output Aggregator]
L --> O
M --> O
N --> O
O --> P[Unified Dashboard]
The architecture resembles a swarm intelligence system: multiple AI agents collaborate while maintaining strict isolation. Each bubble has its own API key, system prompt, and toolset — enabling specialized agents for different programming paradigms.
Launch, monitor, and terminate AI sessions across multiple repositories simultaneously. Each session runs in its own "bubble" with independent memory, file system access, and API quota tracking.
Embed interactive terminals and browser instances directly within sessions. Debug live servers, test API endpoints, or browse documentation without leaving the interface.
Connect to PostgreSQL, MySQL, or SQLite databases directly from any bubble. Run queries, inspect schemas, and generate migration scripts — all orchestrated by AI agents.
Activate specialized modes with a slash command:
/review– Code review with line-by-line suggestions/refactor– Automated code restructuring/deploy– Deployment pipeline management/docs– Documentation generation and update/test– Unit and integration test creation
The interface dynamically adjusts to screen size and task complexity. On a 4K monitor, see 12 bubbles simultaneously. On mobile, collapse to single-bubble focus mode.
Interface and agent prompts available in 15 languages including English, Japanese, German, French, Spanish, Mandarin, Arabic, Russian, Portuguese, Korean, Italian, Dutch, Polish, Turkish, and Vietnamese.
Direct access to maintainers via integrated chat, with average response times under 2 hours during business hours, and under 8 hours during weekends and holidays (2026 support calendar published on wiki).
AeroLaunch acts as a unified API gateway for both Claude and OpenAI models. Configure each bubble to use a different provider or combine them for hybrid workflows:
Example Configuration (aero.config.json):
{
"bubbles": [
{
"id": "frontend-dev",
"provider": "claude",
"model": "claude-opus-4-20260501",
"api_key": "sk-ant-...",
"system_prompt": "Expert React developer with TypeScript focus",
"max_tokens": 64000
},
{
"id": "backend-api",
"provider": "openai",
"model": "gpt-5-turbo-2026",
"api_key": "sk-proj-...",
"system_prompt": "Node.js and Python microservice architect",
"max_tokens": 32000
},
{
"id": "db-optimizer",
"provider": "claude",
"model": "claude-sonnet-4-20260501",
"api_key": "sk-ant-...",
"system_prompt": "Database performance and migration specialist",
"tools": ["sql_console", "schema_analyzer"]
}
],
"global": {
"parallel_limit": 8,
"log_level": "info",
"auto_save_interval": 60
}
}This hybrid approach lets you leverage the strengths of both ecosystems: Claude’s superior reasoning for architecture decisions, and GPT’s speed for boilerplate generation.
Launch AeroLaunch with a custom session profile:
# Basic startup with default profile
aero launch
# Start with three pre-configured bubbles
aero launch --profile full-stack
# Attach to existing session from another terminal
aero attach --session-id "session-2026-03-15-abc123"
# Run a slash command directly
aero exec --bubble "backend-api" "/review --path ./src/services"Sample Console Output:
[2026-03-15 14:22:01] 🛩️ AeroLaunch v2.4.0 initialized
[2026-03-15 14:22:02] 📡 Connecting to Claude API (2 keys active)
[2026-03-15 14:22:02] 📡 Connecting to OpenAI API (1 key active)
[2026-03-15 14:22:03] 🫧 Bubble-1 [frontend-dev] ready (Claude Opus)
[2026-03-15 14:22:03] 🫧 Bubble-2 [backend-api] ready (GPT-5)
[2026-03-15 14:22:03] 🫧 Bubble-3 [db-optimizer] ready (Claude Sonnet)
[2026-03-15 14:22:04] 🔗 Session ID: session-2026-03-15-abc123
> /status
├── Bubble-1: 🟢 Active (4 tasks queued)
├── Bubble-2: 🟡 Loading (package install)
├── Bubble-3: 🟢 Active (schema analysis)
>
Create a reusable profile for your team's workflow (profiles/team-epsilon.yaml):
name: "team-epsilon"
version: "2026.1"
author: "Engineering Team"
description: "Full-stack development profile with 5 specialized agents"
bubbles:
- id: "ui-designer"
provider: claude
model: claude-opus-4-20260501
system_prompt: "Senior UI/UX developer. Expert in Tailwind, Shadcn, and accessibility."
tools: [browser, figma_mcp]
- id: "api-builder"
provider: openai
model: gpt-5-turbo-2026
system_prompt: "REST/GraphQL API designer. Focus on idempotency and rate limiting."
tools: [terminal, curl]
- id: "data-engineer"
provider: claude
model: claude-sonnet-4-20260501
system_prompt: "SQL optimization and ETL pipeline specialist."
tools: [database_console, schema_analyzer]
- id: "tester"
provider: openai
model: gpt-5-turbo-2026
system_prompt: "Test-driven development advocate. Generates Jest, Cypress, Playwright."
tools: [terminal, code_runner]
- id: "reviewer"
provider: claude
model: claude-opus-4-20260501
system_prompt: "Pedantic code reviewer. Security and performance focus."
tools: [code_review, git_analyzer]
settings:
parallel_execution: true
max_threads: 5
log_format: "json"
auto_commit: true
commit_message_template: "[AeroLaunch] {bubble_id}: {summary}"| OS | Version | Status | Notes |
|---|---|---|---|
| 🐧 Linux | Ubuntu 22.04+ | 🟢 Full Support | Native performance |
| 🐧 Linux | Fedora 38+ | 🟢 Full Support | Requires libfuse2 for AppImage |
| 🐧 Linux | Arch (rolling) | 🟢 Full Support | AUR package available |
| 🍎 macOS | 14 (Sonoma) | 🟢 Full Support | ARM and Intel |
| 🍎 macOS | 15 (Sequoia) | 🟢 Full Support | Known issue: Gatekeeper on first launch |
| 🪟 Windows | 10 22H2+ | 🟡 Partial Support | WSL2 required for terminal bubbles |
| 🪟 Windows | 11 23H2+ | 🟢 Full Support | Native Win32 build available |
| 🐚 BSD | FreeBSD 14+ | 🟡 Experimental | Community maintained |
| 📱 iOS | 17+ | 🟡 Limited | SSH client mode only |
| 🤖 Android | 14+ | 🔴 Not Supported | Roadmap for 2027 |
curl -fsSL https://get.aerolaunch.dev/install.sh | bashbrew tap aerolaunch/homebrew-tap
brew install aerolaunchiwr -useb https://get.aerolaunch.dev/install.ps1 | iexdocker pull aerolaunch/server:2026.1
docker run -d -p 8080:8080 -v $(pwd)/config:/config aerolaunch/server:2026.1All sessions automatically persist to disk every 60 seconds (configurable). Resume a session from weeks ago with full context:
aero resume --session-id "session-2026-02-14-valentine-refactor"Enable cross-bubble messaging for collaborative problem-solving:
aero send --from "ui-designer" --to "api-builder" "Need endpoint for user preferences"Built-in telemetry shows real-time API usage, token consumption, and cost tracking:
aero dashboard --metricsOutput:
📊 Session Metrics (session-2026-03-15-abc123)
┌─────────────┬──────────┬──────────┬──────────┐
│ Bubble │ Tokens │ Cost │ Calls │
├─────────────┼──────────┼──────────┼──────────┤
│ frontend │ 142,392 │ $2.13 │ 47 │
│ backend │ 89,441 │ $1.34 │ 31 │
│ database │ 201,883 │ $3.02 │ 68 │
│ tester │ 54,221 │ $0.81 │ 22 │
│ reviewer │ 312,445 │ $4.68 │ 104 │
└─────────────┴──────────┴──────────┴──────────┘
Total: $11.98 | Daily Budget: $50.00 | Remaining: $38.02
Important Legal and Operational Notice:
AeroLaunch is an open-source interface layer for third-party AI APIs. We are not affiliated with Anthropic, OpenAI, or any model provider. Users are responsible for:
- API Key Security – Store keys in environment variables or encrypted vaults. Do not hardcode keys in shared configurations.
- Data Privacy – All data sent to API providers is subject to their respective privacy policies. AeroLaunch does not log or store prompts on our servers.
- Rate Limits & Costs – Parallel execution can rapidly consume API quotas and incur significant costs. Monitor your usage dashboard regularly.
- Compliance – Ensure your use case complies with the terms of service of the underlying API providers, especially regarding automated code generation and deployment.
- No Warranty – This software is provided "as is" without warranty of any kind. The authors are not liable for any damages arising from its use.
As of 2026, this project operates under a strict no telemetry policy. We do not collect usage statistics, crash reports, or any personal information. All operations remain local to your machine unless you explicitly configure remote logging.
This project is licensed under the MIT License – see the LICENSE file for details.
You are free to:
- Use the software for any purpose (commercial or personal)
- Modify and distribute modified versions
- Sub-license under different terms
You must include:
- The original copyright notice
- A disclaimer of liability
We welcome contributions from the community. Please read our Contributing Guidelines before submitting pull requests.
Areas where we especially need help:
- Windows native support improvements
- iOS SSH client integration
- Additional database connectors (MongoDB, Redis, Cassandra)
- Extended language support for UI localization
- Documentation: https://docs.aerolaunch.dev
- Discord Community: https://discord.gg/aerolaunch (24/7 support channel)
- Issues: GitHub Issues (response within 24 hours)
- Email: support@aerolaunch.dev (enterprise customers only)
Built for developers who believe that AI should be a collaborator, not just a tool. AeroLaunch: your distributed intelligence engine for the 2026 era of software development.