Ambient Code Platform
Welcome to the Ambient Code Platform — a Kubernetes-native platform for running AI-powered agents that do real engineering work. Define a task, connect your repos and tools, and let an agent handle the rest.
Get Started
Section titled “Get Started”New to ACP? Start here:
- What is Ambient? — Understand what the platform does and how it works.
- Quickstart (UI) — Create your first agent session through the web interface.
- Core Concepts — Learn about workspaces, sessions, integrations, and workflows.
Dive deeper into the platform’s building blocks:
- Workspaces — Multi-tenant project containers for organizing sessions and teams.
- Sessions — AI agent execution environments with chat, model settings, and artifact output.
- Integrations — Connect GitHub, GitLab, Jira, and Google Workspace.
- Context and Artifacts — Inputs agents work with and outputs they produce.
- Workflows — Structured task templates for common engineering tasks.
Workflows
Section titled “Workflows”Use built-in templates or bring your own:
- Bugfix — Diagnose and fix reported bugs.
- Triage — Classify, prioritize, and route issues.
- Spec-kit — Generate technical specifications from requirements.
- PRD/RFE — Produce product requirements documents.
- Custom Workflows — Load workflows from any git repository.
Extensions
Section titled “Extensions”- GitHub Action — Trigger sessions from CI/CD pipelines.
- MCP Server — Expose ACP capabilities via the Model Context Protocol.
Ecosystem
Section titled “Ecosystem”- Amber — Secret management for agent sessions.
- AgentReady — Prepare repositories for AI agent consumption.