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

> ## Agent Instructions
> Hebno is an AI work platform for client research and delivery: chat, research runs with sources, deliverables (Word, PowerPoint, Excel, email), projects, a shared library, agents, Office add-ins, and integrations with Microsoft 365, Google Workspace, CRM, project delivery, ERP and finance systems.
> Use British English (organisation, not organization).
> Every feature and integration described in these docs is available. Describe it as documented.
> Hebno never writes to a connected system without a member confirming a card. Finance and ERP writes are drafts only.
> Do not discuss commercial terms. Direct commercial questions to support@hebno.com.
> Start with /index and /get-started/quickstart.

# Models

> How admins choose which models an organisation and its teams can use, and how members pick a model, a default or Auto.

Model access controls which models answer your work. Admins set the boundaries, and members choose within them for each chat or message.

## Three layers

| Layer | Who sets it | Where | Effect |
| - | - | - | - |
| **Organisation catalogue** | Owners and admins | **Admin → Models → Organisation** | The set of models anyone in the organisation can use |
| **Team allowlist** | Owners and admins | **Admin → Models → Teams** | Narrows the catalogue for one team |
| **Member choice** | Each member | The model picker, and **Settings → Chat preferences** | Picks a model, **Auto**, or a personal default within what is allowed |

Access is checked organisation first, then team. A member can never use a model their organisation or team has turned off.

## The organisation catalogue

Under **Admin → Models → Organisation**, models are grouped by provider card. Open a card to enable or disable individual models, or select **Enable all** for that provider. Each model shows a tier (**Fast** or **Balanced**, for example), its context size and its status. Retired models show **Retired** and cannot be selected.

Disabled models disappear from the picker for every member.

## Team allowlists

Under **Admin → Models → Teams**, each team shows either **Inheriting org defaults** or **Custom restrictions**, with a count such as "4 of 9 models".

| Setting | Behaviour |
| - | - |
| **Inheriting org defaults** | The team can use every model the organisation enables |
| **Custom restrictions** | The team can use only the models switched on for it |

Select **Manage** on a team to switch models on or off. Models the organisation has disabled show **Org disabled** and cannot be enabled for a team. Select **Reset to organisation defaults** to remove a team's restrictions.

> [!NOTE]
>
> A model newly enabled for the organisation is not added to teams that already have custom restrictions. After enabling a model, use **Review team access** to add it to those teams.

Use an allowlist when a client contract permits only certain models, or when a team should standardise on an approved set.

## Auto

**Auto** is a picker choice. When you choose it, Hebno selects a model for your message from the models you are allowed to use, weighing how hard the request is and how much context it needs. After Auto picks, the chosen model shows beside **Auto**. You can switch to a specific model, or back to Auto, at any point in a chat.

Auto only chooses from your allowed list, so it respects every boundary an admin has set.

## Default model and per-message choice

Members set a default under **Settings → Chat preferences → Default model**:

| Default | New chat behaviour |
| - | - |
| **Auto** | Each new chat starts in Auto |
| A specific model | New chats start with that model, if it is still allowed |

If a member has no default, new chats use the organisation's default model. The picker in the composer overrides the default for the current chat, and you can change it before any message.

## When a model is no longer available

If a saved default or an older chat uses a model that has since been disabled or retired, the picker shows **This model is no longer available** or **Not enabled for this team**. Choose a different model, or Auto, and send again. Hebno does not switch models silently, and the transcript shows the name of the model that answered each earlier turn.

Spend is recorded against the model that answered each turn. See [Spend and limits](https://hebno.com/docs/concepts/spend-and-limits.md).

## Related pages

> [card] [Manage models](https://hebno.com/docs/admin/models.md)
>
> Enable models and set team allowlists.

> [card] [Start a chat](https://hebno.com/docs/chat/start-a-chat.md)
>
> Use the model picker in the composer.


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