> ## Documentation Index
> Fetch the complete documentation index at: https://docs.powabase.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Using the Platform with AI Coding Assistants

> Powabase's REST API is built to be consumed by AI coding assistants like Claude Code, GitHub Copilot, Cursor, and others. How to integrate effectively, plus a preview of the upcoming skill framework.

## Why AI Coding Assistants?

The platform's REST API is the primary interface for building AI applications. A coding assistant can write integration code faster by reading the API structure, generating correct requests, and debugging responses. Every endpoint follows the same patterns for authentication, error format, and streaming, so an assistant with the right context produces working integration code with minimal guidance.

## Getting Started with Claude Code

Claude Code works well with the Powabase API: it can read this documentation, follow the type system, and generate correct API calls in Python, TypeScript, or cURL. For the best results, give Claude Code your project's base URL and API key, and point it at the specific API section you're working with.

<Tip>
  **Share your project context**

  Grab your Project URL and Service Role (Secret) Key from the Studio's Connect modal (click the Connect button in your project header), then tell your coding assistant: "I'm building with the Powabase API. My base URL is \{BASE\_URL} and I'm using the Service Role key for authentication. I need to \[create a knowledge base / build an agent / set up a workflow]." That's enough context for the assistant to generate correct code.
</Tip>

## Common Patterns

The most common tasks AI coding assistants help with when integrating Powabase:

| Task                | What to Ask For                                     | Key API Endpoints                                                           |
| ------------------- | --------------------------------------------------- | --------------------------------------------------------------------------- |
| RAG pipeline setup  | Upload documents, create KB, index, and search      | /api/sources/upload, /api/knowledge-bases, /api/knowledge-bases/{id}/search |
| Agent creation      | Create agent with tools and KB, test with streaming | /api/agents, /api/agents/{id}/tools, /api/agents/{id}/run/stream            |
| SSE stream parsing  | Parse Server-Sent Events in your language           | /api/agents/{id}/run/stream (data: {JSON} format)                           |
| Multi-agent setup   | Create orchestration with entity agents             | /api/orchestrations, /api/orchestrations/{id}/entities                      |
| Workflow automation | Build workflow graph with blocks and edges          | /api/workflows, /api/workflows/{id}/graph                                   |
| Webhook integration | Deploy workflow and set up external trigger         | /api/workflows/{id}/deploy, /api/workflows/{id}/arm                         |

## Upcoming: Skill Framework

We are developing a skill.md framework that gives AI coding assistants structured knowledge of the Powabase API. Instead of relying on general documentation, the skill file provides the assistant with decision trees, common patterns, error handling strategies, and code templates so it can build complete integrations with less hand-holding.

The skill framework will include: API reference in a format suited to LLM consumption, decision logic for choosing indexing strategies and retrieval methods, end-to-end integration templates for common use cases (RAG chatbot, document processing pipeline, multi-agent support team), streaming event parsers for all three languages, and error recovery patterns.

<Info>
  **Coming soon**

  The skill.md framework is under active development. When released, you'll be able to add it to your project's CLAUDE.md or .cursorrules file to give your AI assistant full platform knowledge.
</Info>

## Tips for All AI Assistants

These practices improve results with any AI coding assistant:

| Tip                                              | Why                                                                                     |
| ------------------------------------------------ | --------------------------------------------------------------------------------------- |
| Provide your base URL and API key format         | The assistant can generate ready-to-run code instead of placeholder-filled templates    |
| Reference specific API sections                  | "Use the Knowledge Bases API to..." is more effective than "set up search"              |
| Ask for streaming code in your specific language | SSE parsing differs significantly between Python (requests), TypeScript (fetch), and Go |
| Request error handling                           | The API returns consistent error objects; ask the assistant to handle them              |
| Start with the Quickstart pattern                | Upload → Index → Agent → Stream is the canonical flow                                   |

## Next Steps

<CardGroup cols={2}>
  <Card title="Quickstart" href="/guides/quickstart">
    The canonical RAG agent flow to give your AI assistant as context.
  </Card>

  <Card title="Streaming & SSE" href="/concepts/streaming-patterns">
    Understand the streaming protocol for code generation.
  </Card>

  <Card title="Platform Overview" href="/concepts/platform-overview">
    High-level architecture context for your AI assistant.
  </Card>
</CardGroup>