Skip to content

v0.1.0

Choose a tag to compare

View all tags
@ruslanmv ruslanmv released this 12 Dec 23:40
· 25 commits to master since this release
v0.1.0
78c2b0c
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Release v0.1.0: Initial Release 🚀

Date: December 13, 2025
Package: clouddeploy

We are thrilled to announce the first public release of CloudDeploy, a local-first workspace that solves the "works on my machine" problem by combining your real interactive terminal with an AI deployment copilot.

CloudDeploy runs your actual CLI tools (Docker, IBM Cloud, etc.) in a PTY-backed browser session, while an AI Assistant watches the logs, explains steps, and helps navigate wizard prompts safely.

✨ Key Features in v0.1.0

🖥️ The Web Workspace

  • Real PTY Terminal: Not just fake logs—this is a fully interactive terminal running in your browser (Left Panel).
  • Split-Pane UI: View your terminal output alongside the AI Assistant (Right Panel).
  • Session Management: "Switch Session" picker allows for rapid restarts without killing background processes accidentally.

🤖 AI Copilot & Autopilot

  • Context-Aware Assistance: The AI reads the sanitized tail of your terminal output to explain what is happening in plain English.
  • Autopilot Mode: A guardrailed automation feature that can answer wizard prompts (ENTER, Y/n, numbers) automatically.
  • Error Detection: Flags issues early and suggests the safest next action.

🛡️ Enterprise-Grade Safety

  • Redaction by Default: API keys, Bearer tokens, and passwords are masked before being sent to the LLM.
  • Policy Guardrails: The autopilot and tool layer block destructive commands (e.g., rm -rf, shutdown) and enforce strict input validation.
  • Audit-Friendly: Features a timeline view, summary of steps, and captured issues.

🧰 MCP Tool Server

  • Standard Interface: CloudDeploy implements the Model Context Protocol (MCP).
  • Extensible: Exposes the deployment session as tools (stdio), allowing external agents to observe and reason about the deployment.

🔌 Flexible LLM Support

No vendor lock-in. Configure your preferred provider via environment variables:

  • IBM watsonx.ai (Default/Recommended)
  • OpenAI
  • Anthropic (Claude)
  • Ollama (Local execution)

📦 Installation

CloudDeploy is a Python package. Install it via pip:

pip install clouddeploy

Prerequisites

  • Python: 3.11+
  • OS: macOS / Linux (recommended) or Windows via WSL2.
  • Tools: Ensure your deployment tools (e.g., docker, ibmcloud, jq) are in your system PATH.

🚀 Quick Start

  1. Set your LLM Provider (Example for WatsonX):

    export GITPILOT_PROVIDER=watsonx
export WATSONX_API_KEY="YOUR_KEY"
export WATSONX_PROJECT_ID="YOUR_PROJECT_ID"

  2. Launch the UI:
    Run the UI pointing to your deployment script.

    clouddeploy ui --cmd ./scripts/push_to_code_engine.sh

  3. Open your browser:
    Navigate to http://127.0.0.1:8787 to see your terminal and AI assistant in action.

🧠 v1 Focus: IBM Cloud

While the architecture is provider-agnostic, v0.1.0 includes specific prompt maps and automation heuristics optimized for:

  • IBM Cloud Container Registry
  • IBM Code Engine

Future releases will expand support to multi-cloud providers.

🤝 Contributing & Support

  • Issues: If you encounter edge cases in step detection, please use the "Export Logs" button in the UI and open an issue.
  • Roadmap: We are working on reusable prompt maps, enterprise policy packs, and deeper audit trails.

If CloudDeploy saves you from a deployment incident, please star the repo! ⭐

Assets 2
Loading