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

# Upload documents

> Add knowledge to your AI Experience using an OpenAI Vector Store or xAI Collection, then connect it to Agentsy.

# Upload documents

Your AI Experience can use documents about your expertise, services, methodology, and frequently asked questions to provide more relevant answers.

**Documents are not uploaded directly into Agentsy.** You upload them to a knowledge resource managed by your AI provider—an **OpenAI Vector Store** or an **xAI Collection**—then connect that resource to Agentsy using its ID.

## How it works

1. Use the [Agentsy AI Experience Builder](https://www.agentsy.pro/build-your-ai) to generate your **System Instructions** and **Knowledge Sources Checklist**.
2. Gather the documents you want the AI to reference.
3. Create a knowledge resource with your selected provider and upload the documents there.
4. Copy the resource's ID.
5. In Agentsy, open **Settings → AI Integrations** and paste that ID into the **Vector Store ID** field.
6. Save your settings and test the experience.

> [!NOTE]
>
> The **Vector Store ID** field in Agentsy is used for the provider's knowledge resource. With OpenAI, this is a Vector Store ID. With xAI, the equivalent resource is called a Collection. Use the ID supplied by your provider, and make sure your selected model and integration support that resource.

## 1. Start with your Knowledge Sources Checklist

The [AI Experience Builder](https://www.agentsy.pro/build-your-ai) helps identify which materials would make your experience more useful. Its **Knowledge Sources Checklist** is a plan for what to include; it is **not** itself a knowledge store.

Good starting documents might include:

* An introduction to your expertise or methodology.
* Descriptions of your services, programs, or processes.
* Answers to frequently asked questions.
* Educational guides, articles, or presentations you have permission to use.
* Relevant policies, limitations, or guidance for when visitors should speak with you.

You don't need to upload everything at once. Start with a few accurate, well-organized documents, test the answers, and expand as needed.

> [!TIP]
>
> If you're unsure which documents to prepare, return to the Builder and ask it to prioritize your Knowledge Sources Checklist for a first launch.

## 2. Upload documents with your AI provider

Choose the instructions for the provider you're using in Agentsy.

### OpenAI: Vector Stores

1. Sign in to the [OpenAI Platform](https://platform.openai.com/).
2. Find the area for **Vector Stores** or file-search knowledge resources in your project.
3. Create a Vector Store for your AI Experience.
4. Upload your knowledge files and associate them with that Vector Store.
5. Wait for processing to finish, then confirm the files are available.
6. Copy the **Vector Store ID** provided by OpenAI.

OpenAI's interface may change, so follow the current labels in your platform account. For provider-specific details, see [OpenAI's file search documentation](https://platform.openai.com/docs/guides/tools-file-search).

### xAI (Grok): Collections

1. Sign in to the [xAI Console](https://console.x.ai/).
2. Find the **Collections** area or the provider's current knowledge-management interface.
3. Create a Collection for your AI Experience.
4. Upload the files you want it to use.
5. Confirm that the files have finished processing.
6. Copy the **Collection ID** provided by xAI.

xAI calls this type of searchable knowledge resource a **Collection**, rather than a Vector Store. Check the [xAI documentation](https://docs.x.ai/) for current setup instructions and supported file types.

> [!WARNING]
>
> Do not upload confidential client records, passwords, payment credentials, sensitive personal information, or other restricted material unless you have confirmed that the provider, your permissions, and your data-handling arrangements allow it.

## 3. Connect the knowledge resource in Agentsy

Once the provider has processed your documents:

1. Open your Agentsy workspace.
2. Go to **Settings → AI Integrations**.
3. Locate the **Vector Store ID** field.
4. Paste the ID copied from your OpenAI Vector Store or xAI Collection.
5. Confirm that the selected provider, model, and API key correspond to the resource you created.
6. Save the settings.

| AI provider | Knowledge resource | ID to copy into Agentsy |
| - | - | - |
| OpenAI | Vector Store | Vector Store ID |
| xAI (Grok) | Collection | Collection ID |

**Do not paste** a file name, individual file ID, public document link, or the Knowledge Sources Checklist into the Vector Store ID field. It expects the ID of the knowledge resource.

> [!NOTE]
>
> The resource must be accessible using the provider credentials configured for your Agentsy AI Experience. A resource created under a different project or account may not be available to the API key you're using.

## 4. Test whether your AI can use the documents

After saving the ID, open your AI Experience and ask questions that should be answered by the uploaded files.

For example, if you've uploaded a guide describing your consulting methodology, ask:

* "What are the main steps in your methodology?"
* "How do you approach a new client engagement?"
* "What happens if my situation doesn't fit your usual process?"

Compare the answers with the source documents. Check that the assistant:

* Uses relevant information accurately.
* Doesn't invent facts absent from the documents.
* Acknowledges uncertainty when information is missing.
* Follows the boundaries in its System Instructions.

For a fuller testing process, see [Testing your assistant](/create-and-configure/testing-your-assistant).

## 5. Update your knowledge over time

When your services, methodology, or educational materials change, update the documents **in the provider's knowledge resource**, not in an Agentsy document uploader.

If you keep using the same Vector Store or Collection, you generally don't need to change the ID in Agentsy. If you create a replacement resource, update the **Vector Store ID** field to point to the new one.

Retest the experience after significant changes. If a document contains outdated or conflicting information, remove or replace it in the provider's resource.

## Troubleshooting

| Problem | What to check |
| - | - |
| The assistant doesn't use uploaded information | Confirm processing completed, the resource ID is correct, and the model supports the configured retrieval integration. |
| The Vector Store ID is rejected | Make sure you copied the resource ID—not a file ID—and selected the matching provider. |
| The assistant says it cannot find a document | Confirm the document was added to the connected resource and is in a supported, readable format. |
| The assistant uses outdated information | Check for older or conflicting documents in the provider's knowledge resource. |
| Knowledge works in the provider console but not Agentsy | Check the API key, provider selection, model, resource ID, and access permissions. |

For more help, see [Knowledge retrieval problems](/troubleshooting-and-support/knowledge-retrieval-problems) and [AI models and providers](/create-and-configure/ai-models-and-providers).

## Upload checklist

* [ ] Generate a Knowledge Sources Checklist using the AI Experience Builder.
* [ ] Gather accurate documents you have permission to use.
* [ ] Create an OpenAI Vector Store or xAI Collection.
* [ ] Upload documents and confirm processing completed.
* [ ] Copy the knowledge resource ID.
* [ ] Paste it into **Settings → AI Integrations → Vector Store ID** in Agentsy.
* [ ] Verify the API key and model are compatible with the selected provider.
* [ ] Save and test questions grounded in the uploaded documents.

> [!TIP]
>
> Your documents remain managed in your AI provider's knowledge resource. Agentsy uses the connected resource to support your branded AI Experience.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.