> ## Agent Instructions
> Hebno is a B2B AI control platform for growing teams (30-250 people). Use British English (organisation, not organization).
> Describe only what Hebno supports today. If unsure, say so.
> Spend limits use the organisation display currency in the app.
> Start with /index and /getting-started/quickstart for Hebno orientation.

# Audit trail
Source: https://hebno.com/docs/compliance/audit-trail

What admin actions Hebno records and how operators review activity today.

Hebno records significant admin and budget events server-side.

## What is logged

Examples of recorded actions:

| Area         | Examples                                                       |
| ------------ | -------------------------------------------------------------- |
| **Limits**   | Limit created or updated, threshold approached, spend blocked  |
| **Models**   | Organisation model enabled or disabled, team assignment change |
| **Teams**    | Team created, members assigned                                 |
| **Members**  | Invite sent, member suspended or reactivated, role change      |
| **Settings** | Currency, chat memory, or organisation context updated         |

Each entry includes who made the change and when.

## What you can review in Hebno

Use these views for day-to-day oversight:

* **Admin → Overview** - spend, attribution, and usage trends
* **Admin → Limits** - current caps and spend against each limit
* **Usage** (member menu) - personal spend for the current period

## Detailed audit reports

For a change-history report beyond what admin overview shows, contact [support@hebno.com](mailto:support@hebno.com).

## Next steps

* [Data residency](/docs/compliance/data-residency)
* [Set limits](/docs/guides/set-limits)


# Data residency
Source: https://hebno.com/docs/compliance/data-residency

EU data residency at workspace creation and links to legal documentation.

Hebno is built for teams that need EU-hosted AI processing and clear data handling terms.

## EU data residency

During onboarding **Workspace** setup, you choose whether your organisation requires **EU data residency**.

This records your preference at workspace creation. See [Privacy policy](https://hebno.com/privacy) and [Data processing agreement](https://hebno.com/dpa) for how Hebno describes hosting and processing.

## No training on your data

Hebno does not use customer chat content to train models. This is stated in:

* [hebno.com/#security](https://hebno.com/#security)
* [Privacy policy](https://hebno.com/privacy)
* [Data processing agreement](https://hebno.com/dpa)

## Encryption

| State      | Standard |
| ---------- | -------- |
| At rest    | AES-256  |
| In transit | TLS 1.2+ |

## Subprocessors

Subprocessors are listed in the DPA and privacy documentation at [hebno.com/privacy](https://hebno.com/privacy).

## Access control

Role-based access limits who can change organisation settings, limits, and membership. See [Roles and permissions](/docs/concepts/roles-and-permissions).

## Legal documents

| Document                  | URL                                            |
| ------------------------- | ---------------------------------------------- |
| Privacy policy            | [hebno.com/privacy](https://hebno.com/privacy) |
| Terms of service          | [hebno.com/terms](https://hebno.com/terms)     |
| Data processing agreement | [hebno.com/dpa](https://hebno.com/dpa)         |

For a security overview, see [hebno.com/#security](https://hebno.com/#security).

## What to tell clients

Operators often need a short client-facing summary:

* AI requests for tagged work are attributed in Hebno for reporting
* Data is processed under EU residency commitments in your agreement with Hebno
* Model providers act as subprocessors under the DPA

Direct clients to your executed DPA and privacy terms for contractual detail.

## Next steps

* [Audit trail](/docs/compliance/audit-trail)
* [Overview](/docs/index)


# Attribution
Source: https://hebno.com/docs/concepts/attribution

How team and project tags attach spend to groups and clients, and why tags are sticky on conversations.

Attribution connects each conversation's token cost to the right team and client work. Operators use it for internal chargeback and client rebilling.

## What gets tagged

Every **conversation** can carry:

| Field       | Purpose                                                 |
| ----------- | ------------------------------------------------------- |
| **Team**    | Which group owns the spend                              |
| **Project** | Which client or internal initiative the work belongs to |

Tags are chosen when the conversation is created.

## Sticky tags

Team and project tags are **sticky**. After creation:

* Retagging from the chat body is ignored
* Spend for the whole thread rolls up under the original tags

This keeps monthly reports stable. If attribution was wrong, start a new conversation with the correct tags.

## Projects and clients

Projects belong to the organisation. Each project has:

* A **name** (internal label)
* An optional **client name** (for client-facing reports)

Archived projects cannot be selected for new conversations.

Project tagging is optional. Your organisation can adopt it as internal policy even though Hebno does not require a project on every chat.

## Where attribution appears

Tagged spend flows into:

* **Admin → Overview** charts (by team, by model)
* **Usage** views for members
* **Limit checks** for team-scoped and project-scoped hard limits

## Filtering history

Members filter chat **History** by project from the sidebar. This helps find past client work without mixing unrelated threads.

## Example workflow

An agency delivery lead starts a chat:

1. Select **Team**: Client A delivery
2. Select **Project**: Website copy refresh (client: Client A Ltd)
3. Work in the thread over several days

All token cost for that thread appears under Client A delivery and the Website copy refresh project in admin reporting.

## Next steps

* [Using chat](/docs/guides/chat) - how to tag at conversation start
* [Spend and limits](/docs/concepts/spend-and-limits) - project-scoped limits


# Model access
Source: https://hebno.com/docs/concepts/model-access

How organisations enable models, how teams inherit or restrict access, and member model preferences.

Hebno controls which models members can use through a layered access model: organisation catalogue, optional team restrictions, and personal preferences.

There is no separate policy screen. Access rules live in model enablement and team assignments under **Admin → Models**.

## Organisation level

Under **Admin → Models → Organisation**, admins enable or disable models for the entire workspace.

Disabled models disappear from the model picker for all members unless team-level rules say otherwise.

## Team level

Under **Admin → Models → Teams**, admins configure per-team access:

| Configuration               | Behaviour                                                                                           |
| --------------------------- | --------------------------------------------------------------------------------------------------- |
| **No team assignments**     | Team inherits all organisation-enabled models                                                       |
| **One or more assignments** | **Whitelist mode** - only models marked enabled for that team are available to members on that team |

Team restrictions apply when a member starts or continues a chat tagged to that team.

## Member preferences

Members can set a **default model** under **Settings → Chat preferences** when the picker is available. The default must still be enabled for the organisation and allowed for the member's team context.

Otherwise, new chats follow the organisation default unless the member picks a model in the chat model picker.

## Resolving the active model

When a member sends a message, Hebno resolves the model in this order:

1. Model explicitly chosen in the chat model picker
2. Member default model (if available in settings)
3. Organisation default model

The resolved model must pass organisation and team access checks for the conversation's team tag.

## Gateway

Hebno routes chat requests through its gateway to models enabled in your catalogue. In the app, members choose from model names in the picker - not from a separate provider menu.

## Next steps

* [Manage models](/docs/guides/manage-models) - enable and restrict models in admin
* [Using chat](/docs/guides/chat) - model picker and defaults for members


# Organisations and teams
Source: https://hebno.com/docs/concepts/organisations-and-teams

How workspaces, organisations, teams, and members relate in Hebno.

Hebno uses a multi-tenant workspace model. One person can belong to multiple organisations, each with its own teams, limits, models, and billing.

## Organisation

An **organisation** is your workspace tenant. It holds:

* Members and their roles
* Teams and team membership
* Enabled models and access rules
* Spend limits
* Billing subscription and display currency
* Organisation-wide settings (chat memory policy, organisation context)

The word **workspace** in onboarding refers to the same entity as **organisation**.

## Teams

A **team** groups members for spend attribution and model access. Examples:

* Delivery teams (Design, Engineering)
* Client-facing groups (Client A account team)
* Internal functions (Operations)

Every member can belong to one or more teams. Conversations record the team tag chosen at creation.

Teams do not replace roles. A member on the Marketing team can still be an organisation **admin**.

## Members

A **member** is a person's profile inside one organisation. Each membership has:

* Display name within the organisation
* Role: **owner**, **admin**, or **member**
* Team assignments
* Personal chat preferences (default model, custom instructions, memory toggle)

One global user account can have separate memberships in different organisations.

## Projects

**Projects** sit alongside teams for client work tagging. A project belongs to the organisation and can include an optional client name. Projects are used for:

* Tagging conversations at creation
* Project-scoped spend limits
* Filtering chat history

See [Attribution](/docs/concepts/attribution).

## Typical structure

```
Organisation
├── Teams (Marketing, Client A, …)
├── Projects (Website redesign, Q3 audit, …)
├── Members (each with a role)
└── Limits (org, team, member, or project scope)
```

## Admin vs member experience

| Capability                 | Member | Admin / owner |
| -------------------------- | ------ | ------------- |
| Chat                       | Yes    | Yes           |
| Personal usage view        | Yes    | Yes           |
| Admin overview and limits  | No     | Yes           |
| Invite and suspend members | No     | Yes           |
| Manage models and billing  | No     | Yes           |

See [Roles and permissions](/docs/concepts/roles-and-permissions).

## Next steps

* [Spend and limits](/docs/concepts/spend-and-limits) - how budgets attach to org, team, member, and project
* [Manage members](/docs/guides/manage-members) - invite and suspend teammates


# Roles and permissions
Source: https://hebno.com/docs/concepts/roles-and-permissions

Owner, admin, and member roles and what each can do in chat and admin.

Hebno uses three organisation roles.

## Roles

| Role       | Typical holder            | Summary                                                |
| ---------- | ------------------------- | ------------------------------------------------------ |
| **Owner**  | Founder or account holder | Full admin access; cannot be suspended by others       |
| **Admin**  | Ops lead, IT-light owner  | Full admin access except owner-only billing edge cases |
| **Member** | Individual contributor    | Chat, personal settings, usage view only               |

## Member capabilities

Members with a completed onboarding state can:

* Start and continue chats
* Pick models allowed for their team context
* Attach files to messages
* Tag conversations with team and project at creation
* Browse, rename, and delete their conversation history
* Open personal **Settings**: name, theme, default model, custom instructions, memory (if allowed)
* View **Usage** against limits assigned to them

Members see an **Admin** link in the sidebar but are redirected to chat if they open admin pages without an admin role.

## Admin and owner capabilities

Users with **owner** or **admin** role can access **Admin**:

| Area         | Actions                                            |
| ------------ | -------------------------------------------------- |
| **Overview** | Date-range spend, KPIs, charts, top spenders       |
| **Members**  | Invite, suspend, reactivate (not self, not owner)  |
| **Teams**    | Create teams, assign members                       |
| **Models**   | Enable org models, set team whitelists             |
| **Limits**   | Create and edit hard/soft limits by scope          |
| **Billing**  | View plan, seats, subscription status              |
| **Settings** | Currency, chat memory policy, organisation context |

## Suspension

Admins can **suspend** members. Suspended members cannot use Hebno until reactivated. Owners cannot be suspended. Admins cannot suspend themselves.

## Invitations

Invites specify email, team, and role. Pending invites appear on the members page until accepted or expired.

## Audit events

Admin changes (members, limits, models, settings) are recorded server-side. See [Audit trail](/docs/compliance/audit-trail) for how to review activity in Hebno.

## Next steps

* [Manage members](/docs/guides/manage-members)
* [Organisation settings](/docs/guides/organisation-settings)


# Spend and limits
Source: https://hebno.com/docs/concepts/spend-and-limits

Hard and soft limits, scopes, billing periods, hierarchy rules, and what happens when a cap is reached.

Hebno tracks AI spend and enforces **limits** you set at organisation, team, member, or project level.

## Limits vs subscription

| Concept                  | What it is                                           |
| ------------------------ | ---------------------------------------------------- |
| **Subscription (seats)** | What you pay Hebno monthly per active member         |
| **Limits (AI spend)**    | Internal caps on model usage within a billing period |

Seat billing and AI spend limits are separate. Limits control how much model usage your organisation can incur before requests are blocked or alerts fire.

## Hard vs soft

| Mode     | Behaviour                                                                              |
| -------- | -------------------------------------------------------------------------------------- |
| **Hard** | Blocks new chat requests when spend reaches the limit for the current period           |
| **Soft** | Sends alerts at a threshold and when the limit is reached; does **not** block requests |

Only **hard** limits stop new chat requests. Soft limits are for visibility and operator notification.

When a hard limit blocks a request, the member sees an error and no model call is made.

## Scopes

Set a limit on one of four scopes:

| Scope            | Applies to                             |
| ---------------- | -------------------------------------- |
| **Organisation** | Total spend across the whole workspace |
| **Team**         | Spend attributed to one team           |
| **Member**       | Spend attributed to one person         |
| **Project**      | Spend attributed to one client project |

A single request is checked against every **hard** limit that matches the conversation's organisation, member, team, and project tags.

## Billing period

Limits reset on **UTC calendar months**. The period start is the first day of the month at 00:00 UTC.

Reports and limit checks use the same period boundary.

## Hierarchy rules

When admins create limits, Hebno validates that child budgets do not exceed parents:

* Sum of **team** limits cannot exceed the **organisation** limit
* Sum of **member** limits within a team cannot exceed that **team** limit
* Members not on a team: personal limits cannot exceed the organisation limit
* **Project** and **organisation** scopes skip hierarchy checks against siblings

This prevents configuring caps that contradict each other.

## Alerts

For each limit you can set an **alert threshold** (percentage of the cap). When spend crosses the threshold, admins receive notifications.

At 100% of a soft limit, alerts fire again. At 100% of a hard limit, subsequent requests are blocked.

## Reserve

When viewing limits in admin, **Reserve** is the unallocated budget remaining at organisation or team level after assigned child limits. Use it to see headroom before assigning more caps.

## Example

An agency sets:

* Organisation hard limit: €2,000/month
* Team hard limit (Client A): €800/month
* Member soft limit (freelancer): €100/month with alert at 80%

Client A work tagged to that team counts toward both team and organisation totals. The freelancer gets warnings at €80 but is not blocked unless a hard limit applies to them.

## Next steps

* [Set limits](/docs/guides/set-limits) - admin walkthrough
* [Attribution](/docs/concepts/attribution) - how team and project tags affect spend roll-ups


# Onboarding
Source: https://hebno.com/docs/getting-started/onboarding

How the workspace setup wizard works, including branching for invitations and incomplete organisations.

When you sign in without a ready workspace, Hebno sends you to **Set up workspace**. The wizard creates your organisation, first team, and display currency before you can use chat.

## Step order

The wizard follows this sequence:

| Step            | What you do                                                                                          |
| --------------- | ---------------------------------------------------------------------------------------------------- |
| **Invitations** | Join an existing organisation or choose to create a new workspace (only if you have pending invites) |
| **Profile**     | Enter your full name as it appears to teammates                                                      |
| **Workspace**   | Company name, billing email, country, EU data residency preference                                   |
| **Team**        | Name your first team (creates the organisation and your membership)                                  |
| **Currency**    | Choose display currency for budgets and reports                                                      |
| **Invite**      | Optionally invite teammates by email; you can skip                                                   |

## Where you start

Hebno picks your starting step automatically:

* **Pending invites, no membership** - starts at **Invitations**
* **No invites, no membership** - starts at **Profile**
* **Membership exists but onboarding incomplete** - starts at **Invite** (workspace already created)

## Invitations step

If a colleague invited you:

1. Review pending invitations.
2. **Join** to accept into their organisation, or **Create a new workspace** to start fresh.

Accepting an invite skips workspace and team creation. You inherit the organisation's existing settings.

## Workspace step

| Field             | Purpose                                                      |
| ----------------- | ------------------------------------------------------------ |
| Company name      | Shown in the sidebar and admin shell                         |
| Billing email     | Used for subscription and billing correspondence             |
| Country           | Organisation locale and compliance context                   |
| EU data residency | Records your preference for EU-hosted processing at creation |

## Team step

Creating your first team also creates the organisation and assigns you as **owner**. You can add more teams later under **Admin → Teams**.

## Currency step

Display currency affects how limits and reports are shown. It does not change underlying usage amounts already recorded.

Supported currencies follow your organisation's configuration during onboarding.

## Invite step

Invite teammates by email. You can assign a team and role (member or admin). Skipping this step still completes onboarding.

When onboarding finishes, Hebno redirects you to chat.

## Accepting invites outside the wizard

Invitation links open in Hebno when you are signed in. The page handles:

* Invalid or expired tokens
* Email mismatch (signed-in user differs from invitee)
* Successful accept into the organisation

After accepting, complete any remaining profile steps if required.

## Troubleshooting

| Issue               | What to check                                        |
| ------------------- | ---------------------------------------------------- |
| Stuck in onboarding | Ensure workspace, team, and currency steps completed |
| Invite not visible  | Confirm the invite was sent to your exact email      |
| Wrong organisation  | Sign out and accept the correct invite link          |

See [FAQ](/docs/help/faq) for more common issues.


# Quickstart
Source: https://hebno.com/docs/getting-started/quickstart

Sign up, set up your workspace, send your first chat, and tag spend to a team and project.

This guide gets a new organisation from sign-up to a tagged conversation with spend visibility.

## Before you begin

You need:

* A work email address
* Permission to invite teammates (optional on day one)
* A sense of how you want to split spend (by team, client project, or both)

## Step 1: Create an account

1. Go to [hebno.com/app/signup](https://hebno.com/app/signup).
2. Sign up with your email and password, or use your existing auth provider.
3. If you received an invitation, you can accept it instead of creating a new workspace. See [Onboarding](/docs/getting-started/onboarding).

## Step 2: Complete workspace setup

The onboarding wizard walks you through:

* Your profile name
* Organisation details (company name, billing email, country, EU data residency)
* Your first team
* Display currency for budgets and reports
* Optional teammate invites

When onboarding completes, you land in chat. Details for each step are in [Onboarding](/docs/getting-started/onboarding).

## Step 3: Send your first chat

1. Open **New chat** from the sidebar.
2. Choose a model from the model picker. Only models your organisation has enabled appear.
3. Type a message and send.

<Note>
  If no models appear, an admin must enable models under **Admin → Models**. See
  [Manage models](/docs/guides/manage-models).
</Note>

## Step 4: Tag team and project

When you start a conversation, tag it with:

* **Team** - the group that should own the spend (for example, Marketing or Client A delivery)
* **Project** - optional client or internal work label for attribution

Tags are set when the conversation is created and stay fixed for that thread. See [Attribution](/docs/concepts/attribution).

## Step 5: Check spend

**Members** can open **Usage** from the user menu to see personal spend against any limits assigned to them.

**Admins** open **Admin → Overview** for organisation-wide spend, charts by model and team, and top spenders.

## Next steps

<Steps>
  <Step title="Set limits">
    Configure hard or soft caps at organisation, team, member, or project level.
    [Set limits](/docs/guides/set-limits)
  </Step>

  <Step title="Control model access">
    Enable models for the organisation and restrict by team where needed.
    [Manage models](/docs/guides/manage-models)
  </Step>

  <Step title="Invite members">
    Add teammates and assign roles. [Manage members](/docs/guides/manage-members)
  </Step>
</Steps>


# Billing and seats
Source: https://hebno.com/docs/guides/billing

How Hebno subscription tiers, seats, and AI usage billing work together.

Hebno separates **platform subscription** (seats) from **AI usage** (model cost at pass-through).

## Two cost lines

| Line         | What you pay                 | Controlled by                    |
| ------------ | ---------------------------- | -------------------------------- |
| **Seats**    | Flat per-member subscription | Plan tier and active memberships |
| **AI usage** | Model usage cost from chat   | Usage in chat, capped by limits  |

Hebno does not markup model usage costs. Usage bills at cost, subject to the hard and soft limits you configure.

## Plans and pricing

Starter, Growth, and Scale describe how Hebno is packaged. Current prices and plan details are on [hebno.com/pricing](https://hebno.com/pricing).

## Billing in Hebno

Open **Admin → Billing** to see:

* Current plan tier and status
* Seat count and billing period dates
* Trial state when applicable
* Plan name and renewal dates
* Seat usage versus purchased seats
* Subscription status (active, trialing, past due)

Payment method and invoice details open through the billing portal when available.

Spend overview, limits, model access, and team and project attribution are configured elsewhere in admin.

## Seats

A **seat** is an active organisation membership. Hebno counts active members for billing:

* Starter enforces a minimum of five seats
* Suspended members generally do not count as active

Seat changes sync on your billing period.

## AI usage

Model usage from chat appears in admin spend views and personal **Usage**. Limits under **Admin → Limits** cap how much usage can run in a period.

## Limits vs subscription

Your plan tier does not set AI spend caps automatically. Configure limits separately under **Admin → Limits**.

## Next steps

* [Set limits](/docs/guides/set-limits)
* [Spend and limits](/docs/concepts/spend-and-limits)


# Using chat
Source: https://hebno.com/docs/guides/chat

Start conversations, pick models, tag team and project, attach files, and manage history.

Chat is the primary member experience in Hebno. Every message runs through models your organisation enables, with spend recorded automatically.

## Open chat

After onboarding, open **New chat** from the sidebar.

The chat panel includes:

* Message input
* Model picker
* Team and project tag selectors (at conversation start)
* File attachment support (drag-and-drop)

## Choose a model

Use the model picker to select which model answers this thread. Only models enabled for your organisation and allowed for the conversation's team appear.

If the default model control is available, set it under **Settings → Chat preferences → Defaults**. Otherwise Hebno uses the organisation default unless you pick a model per chat.

## Tag team and project

When starting a conversation, select:

1. **Team** - required for correct spend roll-up in most setups
2. **Project** - optional; use for client or initiative attribution

Tags cannot be changed after creation. Start a new chat if attribution was wrong. See [Attribution](/docs/concepts/attribution).

## Attach files

Drag files into the chat panel or use the attachment control. Files upload to private organisation storage.

## Regenerate responses

Use **Regenerate** on an assistant message to request a new answer from the same model. Each regeneration creates additional usage and cost.

## History

Past conversations appear in **History**, grouped by date. You can:

* Search and open threads
* Rename conversations
* Delete one or many via the manage dialog
* Filter by project from the sidebar

## Personal settings

Open **Settings** from the user menu:

| Section              | Options                                                     |
| -------------------- | ----------------------------------------------------------- |
| **Personal**         | Display name, theme                                         |
| **Chat preferences** | Default model, custom instructions                          |
| **Memory**           | Personal memory toggle (only if organisation allows memory) |

**Organisation context** (background about your company injected into every chat) is set by admins under **Admin → Settings**, not by members.

## Usage

Open **Usage** from the user menu to see personal spend for the current period against any member-scoped limits.

## Organisation context and memory

Admins may set **organisation context** and **chat memory** policy under **Admin → Settings**. When memory is allowed, members can turn on **Remember things about me** under **Settings → Chat preferences → Memory**.

## Next steps

* [Model access](/docs/concepts/model-access)
* [Set limits](/docs/guides/set-limits) - if you are blocked by a limit


# Manage members
Source: https://hebno.com/docs/guides/manage-members

Invite teammates, assign roles and teams, and suspend or reactivate members.

Member management is under **Admin → Members**. Requires **admin** or **owner** role.

## View members

The members page lists:

* Active members with role and team assignments
* Pending invitations (email, role, invited date)

## Invite a member

1. Click **Add member**.
2. Enter email address.
3. Choose **role**: member or admin.
4. Assign a **team** (recommended for attribution and model access).
5. Send invitation.

The invitee receives an email with a link to accept. They must sign in with the invited email to accept.

## Roles when inviting

| Role       | When to use                                      |
| ---------- | ------------------------------------------------ |
| **Member** | Default for contributors who only need chat      |
| **Admin**  | Ops leads who manage limits, models, and billing |

Organisation **owner** is assigned at workspace creation, not via invite role picker.

## Suspend a member

Admins can suspend members who should lose access immediately.

* Suspended members cannot sign in to the product
* You cannot suspend yourself
* You cannot suspend the **owner**

Reactivate suspended members from the same members page.

## Pending invites

Cancel or resend pending invites from the members list. Expired tokens show an error on the invite acceptance page.

## Seats and billing

Each active membership counts toward **seats** billed on your subscription. Suspended members typically do not count as active seats (see [Billing](/docs/guides/billing)).

Starter plan has a minimum of five seats.

## Audit

Invite, suspend, reactivate, and role changes write audit events (`user.*` namespace).

## Next steps

* [Roles and permissions](/docs/concepts/roles-and-permissions)
* [Onboarding](/docs/getting-started/onboarding) - invite step in the wizard


# Manage models
Source: https://hebno.com/docs/guides/manage-models

Enable models for your organisation and restrict access by team with whitelist rules.

Model management lives under **Admin → Models** with two tabs: **Organisation** and **Teams**.

## Organisation tab

Enable or disable models for the entire workspace.

| Action      | Effect                                                                        |
| ----------- | ----------------------------------------------------------------------------- |
| **Enable**  | Model appears in pickers for members whose team inherits org access           |
| **Disable** | Model hidden organisation-wide unless a team whitelist explicitly includes it |

Start here when rolling out a new model or retiring one from general use.

## Teams tab

Configure per-team access for members tagged to that team in chat.

| Team configuration   | Member experience                                        |
| -------------------- | -------------------------------------------------------- |
| No rows for the team | Inherits all organisation-enabled models                 |
| One or more rows     | **Whitelist** - only models marked enabled for that team |

Use team whitelists when:

* A client contract allows only specific models
* A junior team should not access premium models
* A delivery pod should standardise on one approved model

## Workflow: roll out a new model

<Steps>
  <Step title="Enable at organisation">
    Turn the model on under **Organisation** so it enters the org pool.
  </Step>

  <Step title="Restrict sensitive teams">
    Under **Teams**, whitelist only approved models for restricted teams.
  </Step>

  <Step title="Communicate">
    Tell members to start new chats with the correct team tag so whitelist rules
    apply.
  </Step>
</Steps>

## Workflow: remove a model

1. Disable the model under **Organisation**.
2. Remove or disable team whitelist rows that reference it.
3. Confirm members can finish existing threads; new chats will not offer the model.

## Next steps

* [Model access](/docs/concepts/model-access)
* [Using chat](/docs/guides/chat)


# Organisation settings
Source: https://hebno.com/docs/guides/organisation-settings

Configure display currency, chat memory policy, and organisation context for all member chats.

Organisation settings are under **Admin → Settings**. Only **admin** and **owner** roles can change them.

## Display currency

**Display currency** controls how limits, dashboards, and usage summaries are formatted for your organisation.

Changing currency affects display only. Historical usage amounts are unchanged.

Currency can also be set during onboarding.

## Chat memory

**Chat memory** is an organisation-wide policy toggle:

| Setting        | Effect                                                 |
| -------------- | ------------------------------------------------------ |
| **Allowed**    | Members can enable personal memory in their settings   |
| **Disallowed** | Personal memory toggles are hidden; no extraction runs |

When allowed, members choose whether memory is on for their account under **Settings → Chat preferences → Memory**.

## Organisation context

**Organisation context** is background text admins provide about the company (services, tone, key facts). Hebno prepends it to the system prompt for **all** member chats in the organisation.

Use it for stable facts you want every model to know:

* What your firm does
* Standard client terminology
* Compliance reminders that apply to every thread

Keep it concise. Long context consumes tokens on every request.

## Next steps

* [Using chat](/docs/guides/chat) - how members experience context and memory
* [Manage models](/docs/guides/manage-models)


# Set limits
Source: https://hebno.com/docs/guides/set-limits

Create hard and soft spend limits at organisation, team, member, and project scope in Admin.

Limits cap AI usage spend within each UTC calendar month. Configure them under **Admin → Limits**.

## Open Limits

1. Sign in as **admin** or **owner**.
2. Open **Admin** from the sidebar.
3. Select **Limits**.

The page lists existing limits with current spend versus cap, enforcement mode, and scope.

## Create a limit

1. Click **Set limit**.
2. Choose **scope**: organisation, team, member, or project.
3. Select the target entity (team, member, or project) when scope is not organisation.
4. Enter the limit amount in your organisation's **display currency**.
5. Choose enforcement:
   * **Hard** - block requests when the cap is reached
   * **Soft** - alert only
6. Set **alert threshold** percentage (for example, 80% to warn before the cap).
7. Save.

Hebno validates hierarchy rules when you save team or member limits. If child limits would exceed a parent, the save is rejected with a clear error.

## Hard limit behaviour

When spend for the period meets or exceeds a hard limit:

* New chat requests are blocked with an error
* No model call is made
* Admins receive notifications

Members already mid-stream may complete an in-flight response; the block applies to the next request.

## Soft limit behaviour

Soft limits never block. They trigger notifications at the alert threshold and again at 100%.

Use soft limits for visibility on individuals or projects where you want warnings before considering a hard cap.

## Practical patterns

| Scenario                 | Suggested setup                  |
| ------------------------ | -------------------------------- |
| Agency monthly AI budget | Organisation hard limit          |
| Client delivery cap      | Team hard limit per client team  |
| Junior staff guardrail   | Member soft limit with 80% alert |
| Fixed-fee project        | Project hard limit               |

## Reserve

After assigning team and member limits, check **Reserve** on organisation and team rows. Reserve is headroom not yet allocated to child limits.

## Edit and remove

Admins can update limit amounts, enforcement mode, and thresholds. Changes apply to the current period's checks immediately.

Removing a limit stops future enforcement for that scope; past usage is unchanged.

## Next steps

* [Spend and limits](/docs/concepts/spend-and-limits) - periods, hierarchy
* [FAQ](/docs/help/faq) - blocked by limit troubleshooting


# FAQ and troubleshooting
Source: https://hebno.com/docs/help/faq

Common questions about limits, invites, models, currency, and access errors.

## I was blocked from sending a chat message

**Symptom:** Error when sending a message.

**Cause:** A **hard limit** for the current UTC month was reached at organisation, team, member, or project scope.

**Fix:**

1. Ask an admin to check **Admin → Limits** for caps at or above 100%.
2. Raise the limit or wait until the next UTC month.
3. Confirm the conversation's team and project tags - they determine which limits apply.

Soft limits do not block. See [Spend and limits](/docs/concepts/spend-and-limits).

***

## A model does not appear in the picker

**Possible causes:**

| Cause                             | Fix                                                                   |
| --------------------------------- | --------------------------------------------------------------------- |
| Model disabled for organisation   | Admin enables it under **Models → Organisation**                      |
| Team whitelist excludes the model | Admin updates **Models → Teams** or member retags team on a new chat  |
| Default model not available       | Member picks model per chat or checks **Settings → Chat preferences** |

See [Model access](/docs/concepts/model-access).

***

## My invite link does not work

| Symptom        | Fix                                    |
| -------------- | -------------------------------------- |
| Expired token  | Admin resends invite from **Members**  |
| Email mismatch | Sign in with the invited email address |
| Invalid token  | Request a new invitation               |

Accept the invitation while signed in with the invited email address.

***

## Spend shows in the wrong team or project

Tags are set at **conversation creation** and cannot be changed. Start a **new chat** with the correct team and project tags.

Past usage remains under the original tags.

***

## Currency display looks wrong

Display currency is set in onboarding or **Admin → Settings**. If amounts look wrong after a change, allow time for dashboards to refresh.

***

## I see Admin in the sidebar but get redirected

Only **owner** and **admin** roles access admin pages. Members are redirected to chat.

Ask an organisation admin to upgrade your role if you need admin access.

***

## Where are audit logs?

Spend and usage are in **Admin → Overview** and **Usage**. For change history and audit reporting, see [Audit trail](/docs/compliance/audit-trail) or contact [support@hebno.com](mailto:support@hebno.com).

***

## Is there an API?

Hebno does not offer a public API at this time. Contact [support@hebno.com](mailto:support@hebno.com) if you need integration guidance.

***

## Still stuck?

Contact [support@hebno.com](mailto:support@hebno.com) or [book a demo](https://hebno.com/demo) for a walkthrough with the Hebno team.


# For AI agents
Source: https://hebno.com/docs/help/for-ai-agents

How coding agents and LLM tools should discover, install, and query Hebno product documentation.

This page is for AI assistants, coding agents, and automation tools that need accurate Hebno product context.

## Start here

| Resource               | URL                                                                                                  | Purpose                                     |
| ---------------------- | ---------------------------------------------------------------------------------------------------- | ------------------------------------------- |
| **Curated index**      | [hebno.com/docs/llms.txt](https://hebno.com/docs/llms.txt)                                           | Priority-ordered map of documentation pages |
| **Capability summary** | [hebno.com/docs/skill.md](https://hebno.com/docs/skill.md)                                           | What Hebno enables, inputs, and constraints |
| **Overview**           | [hebno.com/docs](https://hebno.com/docs)                                                             | Hebno overview for humans and agents        |
| **Quickstart**         | [hebno.com/docs/getting-started/quickstart.md](https://hebno.com/docs/getting-started/quickstart.md) | First-run workflow                          |

Append `.md` to any doc path to fetch Markdown instead of HTML.

## Install the Hebno skill

Install the hosted skill so your agent loads Hebno capabilities before answering product questions:

<Prompt description="Install the Hebno documentation skill for AI coding tools.">
  npx skills add [https://hebno.com/docs](https://hebno.com/docs)
</Prompt>

The skill describes tasks Hebno supports, required inputs (organisation, team, project, limits), and guardrails (hard limits block chat, no public API).

## Search MCP server

Mintlify hosts a read-only **search MCP** on public documentation:

```
https://hebno.com/docs/mcp
```

Connect this URL in Cursor, Claude, or other MCP clients to search live documentation content.

Use the page header contextual menu (**Add MCP**, **Cursor**) on any doc page to copy install commands.

<Note>
  This MCP searches published documentation only. It does not administer your
  Hebno organisation or call product APIs.
</Note>

## Agent instructions (site-wide)

Every Markdown export includes these rules from `docs.json`:

* Hebno is B2B AI control for teams of 30-250 people
* Use British English (`organisation`, not `organization`)
* Describe only what Hebno supports today. If unsure, say so.
* Spend limits use the organisation display currency in the app

## Recommended reading order

1. [Overview](/docs/index)
2. [Spend and limits](/docs/concepts/spend-and-limits)
3. [Roles and permissions](/docs/concepts/roles-and-permissions)
4. [Set limits](/docs/guides/set-limits) or [Using chat](/docs/guides/chat) depending on the question

## llms-full.txt

Mintlify auto-generates [hebno.com/docs/llms-full.txt](https://hebno.com/docs/llms-full.txt) with full site content for bulk indexing. Prefer `llms.txt` for navigation; use `llms-full.txt` only when you need entire-corpus context.

## Do not assume

| Topic                  | Guidance                                                                                |
| ---------------------- | --------------------------------------------------------------------------------------- |
| Public customer API    | [FAQ](/docs/help/faq)                                                                        |
| Detailed audit reports | [Audit trail](/docs/compliance/audit-trail) or [support@hebno.com](mailto:support@hebno.com) |

Describe only what Hebno supports today. If unsure, say so and point to the docs or support - do not infer from hebno.com/pricing alone.


# Overview
Source: https://hebno.com/docs/index

What Hebno is, who it is for, and the three pillars of AI control for growing teams.

Hebno is an AI control platform for growing teams. It gives operators one place to access every major model their organisation uses, track spend by team and client, set hard budgets, and keep processing inside the EU.

## Who Hebno is for

Hebno is built for non-technical operators at agencies, consultancies, and professional services firms with 30-250 people. The typical buyer is a CEO, COO, or Head of Ops who owns the budget and needs visibility without deploying infrastructure.

Hebno is not built for developers running their own gateways, or for enterprise IT departments buying shelfware.

## Three pillars

<Columns>
  <Card title="Multi-model access" icon="layers">
    One chat interface to the models your organisation enables. Assign access by
    team and role without asking members to manage separate accounts.
  </Card>

  <Card title="Spend control" icon="chart-line">
    Live cost by person, team, and client project. Set hard monthly limits that
    block spend when a cap is reached, or soft limits that alert before you hit
    the ceiling.
  </Card>

  <Card title="EU compliance" icon="shield-check">
    EU data residency at workspace creation. No training on your data. Audit
    events recorded for admin actions.
  </Card>
</Columns>

## How billing works

Hebno charges for **seats**, not tokens. AI usage is billed separately at cost with no markup, capped by the limits you set.

See [Billing and seats](/docs/guides/billing) for plan tiers and how seats relate to active members.

## Get started

<CardGroup>
  <Card title="Quickstart" icon="rocket" href="/getting-started/quickstart">
    Sign up, complete onboarding, and send your first chat in under ten minutes.
  </Card>

  <Card title="Core concepts" icon="book-open" href="/concepts/spend-and-limits">
    Understand organisations, limits, model access, and attribution before you
    configure controls.
  </Card>
</CardGroup>

Open Hebno at [hebno.com/app](https://hebno.com/app). For a guided walkthrough, [book a demo](https://hebno.com/demo).