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

# Knowledge (RAG collections)

> List and delete knowledge collections for RAG in workflows and agents

## What is Knowledge (RAG)?

**Knowledge** in Flowra is your project’s **knowledge base** that AI agents and workflows can use to answer questions accurately instead of guessing. The technology behind it is called **RAG** (Retrieval-Augmented Generation): the system stores your documents (PDFs, text, web pages, etc.) in small pieces called **chunks**, and when a user asks something, it **finds the relevant chunks** and sends them to the AI so the answer is based on *your* content, not only the model’s training.

Without a knowledge base, the AI only has what you type in the chat and its general training. With a knowledge base, it can say things like: *“According to your pricing doc, the Pro plan is \$29/month”* or *“Your refund policy states that…”* — because it’s reading from the documents you’ve added.

### Example use cases

* **Support bot** — Add your FAQ, help articles, and policy docs into a collection. When a customer asks “What’s your return policy?”, the agent finds the right paragraph and answers from your docs.
* **Internal assistant** — Upload company handbooks, process docs, or project briefs. Employees ask the agent questions and get answers grounded in your internal knowledge.
* **Product docs** — Feed the agent your product documentation so it can answer “How do I enable 2FA?” or “What’s the rate limit for the API?” using your own wording.

### How it works in Flowra

1. You **create a knowledge collection** (in the [Flowra Dashboard](https://flowra.dev) or in chat) and give it a name (e.g. “Support FAQ”, “Product docs”).
2. You **add content** to it: upload files, paste text, or add URLs. Flowra splits them into chunks and indexes them.
3. When you **create or edit a workflow/agent**, you attach one or more collections via **allowedKnowledgeCollectionIds**. From then on, when that agent runs, it can search those collections and use the retrieved chunks to generate answers.

Creating and editing collections (adding/removing documents) is done in the product. The **API** lets you **list** all collections and **delete** a collection (and all its chunks) so you can manage them from your own app or scripts.

***

## API endpoints

### Get all RAG collections

**GET** `/api/v1/knowledge/collections` returns all knowledge collections for the current project. Each item includes:

| Field        | Description                                                                                        |
| ------------ | -------------------------------------------------------------------------------------------------- |
| `id`         | Collection ID — use for `allowedKnowledgeCollectionIds` when creating or updating agents/workflows |
| `name`       | Collection name                                                                                    |
| `chunkCount` | Number of chunks in the collection (optional)                                                      |

<CodeGroup>
  ```bash theme={null}
  curl -X GET "https://flowra.dev/api/v1/knowledge/collections" \
    -H "x-api-key: YOUR_API_KEY"
  ```
</CodeGroup>

### Delete a knowledge collection

**DELETE** `/api/v1/knowledge/collections/{id}` deletes a knowledge collection and all its chunks. Use the collection `id` from **GET** `/api/v1/knowledge/collections`.

<CodeGroup>
  ```bash theme={null}
  curl -X DELETE "https://flowra.dev/api/v1/knowledge/collections/755ab29c-2864-5cfd-9c2b-4953f7314b21" \
    -H "x-api-key: YOUR_API_KEY"
  ```
</CodeGroup>

## Related

* [Workflows](/guides/workflows) — Attach knowledge collections to workflows (e.g. `allowedKnowledgeCollectionIds`).
* [Files](/guides/files) — Upload files that can be used in knowledge bases.
* [API reference](/api-reference) — Full request/response schemas for knowledge endpoints.