AI Library Coding Agent – VS Code Extension

Build, automate, and orchestrate workflows directly inside VS Code using AI Library’s agentic platform.

AI Library brings enterprise-grade AI agents into your IDE — enabling you to build faster, integrate seamlessly, and automate complex workflows with a fully managed backend.

🚀 What is AI Library?

AI Library is an agentic automation platform designed for production-grade AI deployments. The Coding Agent extension connects your IDE to AI Library’s managed infrastructure, giving you access to tools, memory, orchestration, and integrations — without complex setup.

🧠 Agentic Development

• Use AI Library agents and tools without manually referencing documentation
• Agents can
  • Read, write, and edit files
  • Run bash commands
  • Search the web
  • Call external tools and APIs
  • Orchestrate multi-step workflows

🧰 Documented Skills

The extension can invoke a set of documented AI Library “skills” (tool APIs). Skill docs live in ai-library-plugin/skills.

• Agent Management: create-agent, get-agent, update-agent
• Chat: chat-agent (chat with an agent; can return structured JSON when using the agent’s form schema)
• Notes (Lightweight DB, CRUD): create/get/list/update/delete notes attached to a resource (e.g., an agent)
• Agent Knowledge Base: agent-knowledge-add, agent-knowledge-list, agent-knowledge-search, delete-knowledge-sources
• Forms (Structured Output): update-form (update a form schema by form_id)
• Files & Document Parsing: upload-files (upload local files → hosted URLs), read-documents (extract text from doc URLs)
• Speech / Audio: transcribe (audio → text), generate-speech (text → audio URL)
• Web: websearch (web or image search with optional filters)

🔄 Session Continuity

• Persistent session memory across messages
• Context-aware development assistance
• Ability to start a new session or clear history anytime

⚙️ Configurable Execution

• Select model dynamically
• Configure permission modes
• Define allowed tools and execution scope
• Control agent autonomy vs. confirmation prompts

📡 Live Tool Streaming

• Stream tool calls and execution results in real time
• Full visibility into what the agent is doing
• Inspect commands before execution (depending on permission mode)

☁️ Fully Managed Backend

• No infrastructure to manage
• Scalable agent runtime
• Secure execution and orchestration layer
• Reduced time to development

🏗️ How It Works

1. Install the extension
2. Authenticate using your AI Library API key
3. Start a session from the command palette
4. Prompt the agent directly from the IDE
5. Watch it reason, call tools, and modify your workspace in real time

🔐 Security

• API keys are stored securely using VS Code’s built-in Secret Storage
• Keys are never stored in settings.json
• Fine-grained permission controls for file system and shell access
• bypassPermissions mode should only be used in trusted environments

🧩 Ideal Use Cases

• Rapid backend scaffolding
• Infrastructure automation (Terraform, Docker, AWS, etc.)
• API integrations
• Refactoring large codebases
• Documentation generation
• Multi-step developer workflows

🛠️ Configuration Options

Coming Soon

• Workspace-level agent collaboration
• Multi-agent orchestration
• Deep CI/CD integration
• Enterprise SSO and team controls