> For the complete documentation index, see [llms.txt](https://shoppad.gitbook.io/yedric/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://shoppad.gitbook.io/yedric/going-further/organizations.md).

# Organizations

Organizations serve as the top-level structure within yedric.ai, forming the foundation of how your entire environment is arranged and managed. Every element of the platform, from Assistants down to individual Users, exists within the context of an Organization, making it the central point of control for access, structure, and oversight.

## Organization → Assistants → Members

Organizations sit at the top of the hierarchy. Within an Organization, Assistants are configured to carry out specific functions of your AI interfaces, and Members (your team) are given access to manage them.

This structure ensures that access and functionality are organized in a deliberate, manageable way.

> **Note:** "Members" here refers to your internal team, the people who build and manage your assistants. This is separate from the end-users who interact with your Yedric widget.

## When to Use Organizations

If you are working solo or just getting started, you may not need to configure anything here right away, since Yedric works without any organization setup. Organizations become relevant when you want to invite teammates to help manage your assistants, control who can edit assistants or change settings through roles, or keep multiple assistants organized under a shared workspace.

## Accessing and Managing Organizations

Open **Organization** in the left sidebar within [app.yedric.ai](http://app.yedric.ai). It has sections for **Settings**, **Live Chat**, **AI Service Keys**, **Developer usernames**, **Secrets**, **Integrations**, and **Usage**.

### Inviting Members

To expand access to your Organization, members can be invited via email. When sending an invitation, you will assign one of three roles:

* **Owner:** Full control over the organization, including the ability to manage billing and delete the organization. Typically the person who created it.
* **Admin:** Elevated access for those responsible for managing the organization, its assistants, and its members. Admins can invite and remove members, manage API keys and secrets, and configure organization settings.
* **Member:** Standard access for users who work within the organization. Members can view and edit assistants but cannot manage organization-level settings or billing.

## Developer Usernames

If members of your team interact with your assistants for testing or development purposes, their conversations will appear alongside real user conversations in the Conversations tab. Developer Usernames lets you map team members' accounts to display names so their conversations are automatically labeled as internal and kept separate.

To set this up, open **Organization > Developer usernames** and add the accounts you want to tag.

***

## Settings

The **Settings** section holds the organization's own name, the one shown in the organization switcher. Renaming it does not affect your assistants or their widgets.

***

## Live Chat

The **Live Chat** section holds one organization-wide control: a global limit on how many active chats a single teammate can hold at once, counted across every assistant. When it is set, it overrides the per-assistant **Max concurrent chats per agent** field, and that field says so. Leave it unset to let each assistant decide for itself. See [Live Chat](/yedric/going-further/live-chat.md).

***

## AI Service Keys

Your assistants run on large language models from Google, OpenAI, and Anthropic. The **AI Service Keys** section is where you add your own provider API keys so the platform can call those models on your behalf. Keys are set once per organization and shared by every assistant in it.

You can get started without adding anything, since GPT-4o Mini works with no key at all. To use any other model, add the matching provider's key on this page.

| Provider         | Where to get it                                                                    |
| ---------------- | ---------------------------------------------------------------------------------- |
| Google AI Studio | [aistudio.google.com/app/apikey](https://aistudio.google.com/app/apikey)           |
| OpenAI           | [platform.openai.com/api-keys](https://platform.openai.com/api-keys)               |
| Anthropic        | [console.anthropic.com/settings/keys](https://console.anthropic.com/settings/keys) |

The keys you add here also determine which models appear in the Model dropdown when you create or edit an assistant. If a model you want is not listed, check that the matching provider key is set.

Each saved key is masked. Click **Reveal** to display the stored value on demand, or **Replace** to enter a new one. If an assistant is set to a model whose provider key is missing, its chat will not return a response until you add the key here.

> **Note:** Managing AI service keys requires the Owner or Admin role.

***

## Secrets

**Secrets** is an encrypted store for the API keys and tokens your custom tools ([Actions](/yedric/going-further/actions.md)) need to authenticate with outside services: a Stripe key, a Shopify token, an internal API bearer, and so on. Storing them here keeps credentials out of your assistant's prompt and out of the tool configuration itself.

Each secret is a **Name** and a **Value**. To add one, enter a descriptive name (e.g. `Stripe API Key`) and its value (e.g. `sk_live_…`), then click **Add**. Saved secrets are listed with their creation date and three actions:

* **Reveal / Hide**: fetch and display the decrypted value on demand. It is never held in the dashboard longer than the row is revealed.
* **Copy**: copy the revealed value to your clipboard.
* **Delete**: permanently remove the secret.

Values are encrypted at rest. To use a secret in an Action's authentication settings, reference it by name with `{{secret:Name}}`. The platform substitutes the decrypted value at request time so the raw credential never appears in your configuration.

These are organization-level secrets, available to every assistant in the organization. When a single assistant needs its own private credential, use assistant-level secrets from that assistant's **Settings** tab instead; those take precedence over an organization secret with the same name.

> **Note:** Managing secrets requires the Owner or Admin role.

***

## Integrations

**Integrations** is where you connect Yedric to the systems your team already uses. It holds two kinds of app:

* **Custom App** renders your own content in a panel beside a conversation, so your team can see a customer's account details without leaving Yedric.
* **Webhooks** post events to a URL you control as they happen, so your own systems can react when a conversation is created, escalated, or summarized.

Both are set up per organization and shared by everyone in it. For the full walkthrough, see [Integrations](/yedric/going-further/integrations.md).

> **Note:** Integrations require the Owner or Admin role, to read as well as to change.

***

## Usage

The **Usage** section reports how much your organization's assistants have consumed, so you can keep an eye on volume across every assistant at once. Choose a window of **7 Days**, **30 Days**, or **90 Days**.

Five totals summarize the selected window:

* **Total Requests**: the number of model requests made.
* **Prompt Tokens**: tokens sent to the models (your prompts, context, and history).
* **Completion Tokens**: tokens generated by the models in their responses.
* **Total Tokens**: prompt and completion tokens combined.
* **Cached Prompt Tokens**: for models that support prompt caching, the number of prompt tokens served from cache rather than re-processed, with the cache hit rate shown beneath it as a percentage of prompt tokens. A higher hit rate means faster responses.

Below the totals, a **By Model** table breaks the same numbers down per model, so you can see which models are driving your usage. **By Assistant** and **By Username** tables do the same per assistant and per verified end-user username. If there is no activity in the selected window, the section shows "No usage data yet."

Each individual assistant also has its own Usage page in the assistant's sidebar, showing the same metrics scoped to just that assistant.