> For the complete documentation index, see [llms.txt](https://docs.ovaledge.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ovaledge.com/release8.3/askedgi/user-guides/consumption-and-usage-guide.md).

# Consumption and Usage Guide

Consumption and usage within **askEdgi** define how resources, compute, and AI-powered capabilities are utilized across different deployment models. This document provides a comprehensive understanding of the **spend behavior**, **billing triggers**, and **usage visibility** for users in both **Public** and **Enterprise (SaaS)** editions.

askEdgi operates on a pay-as-you-consume principle, meaning every AI interaction, workspace initiation, file upload, or recipe execution contributes to a user’s defined spend limit.

{% hint style="info" %}
To understand the terminology used across askEdgi, refer to the askEdgi Glossary available on the support portal [here](https://galaxy-new.ovaledge.net/#nav/glossary-list?browse=list).
{% endhint %}

## Overview of Consumption Models

askEdgi offers two primary consumption and billing models:

<table><thead><tr><th width="219.333251953125">Edition</th><th>Description</th></tr></thead><tbody><tr><td>Public Edition</td><td>Self-service model with pre-defined spend limits, trial periods, and public recipe monetization. Ideal for individuals, freelancers, and learners exploring AI-driven analytics.</td></tr><tr><td>Enterprise (SaaS) Edition</td><td>Subscription-based enterprise deployment with controlled OpenAI API provisioning, monitored budgets, and centralized billing. Designed for governed organizational usage.</td></tr></tbody></table>

Each model differs in how spend limits are enforced, how billing begins, and how resource consumption is tracked. The following sections describe these differences in detail.

## Enterprise (SaaS) Edition Consumption and Usage

### Subscription and Included Quota

In the Enterprise model, consumption is governed under a controlled SaaS environment. Each organization receives an **allocated monthly AI usage quota** provisioned via OvalEdge’s integrated OpenAI account.

* By default, each enterprise tenant is allocated **$100 per month** worth of OpenAI API usage.
* Beyond this, additional calls are billed at **$0.02 per API request**, managed by OvalEdge.
* Alternatively, enterprises can configure their own **OpenAI API key** for direct billing and quota management.

### Billing Model and Quota Enforcement

OvalEdge enforces spend and usage limits using OpenAI’s built-in project-level controls:

* **Budget Threshold Alerts**: Trigger notifications at 80% of quota consumption.
* **Budget Hard Caps**: Automatically halt API calls once the maximum monthly budget is reached.
* **Rate Limiting**: Restricts excessive usage to prevent performance degradation.
* **Model Access Control**: Ensures only authorized models are used per plan tier.

Billing is consolidated at the organizational level, enabling unified visibility and estimates of future spend based on current and historical usage through enterprise dashboards and the My Usage interface.

Administrators can also view actual AI token consumption in the **AI Consumption Dashboard** under A**udit Trails > AI.** That dashboard reports consumption only. It does not show billing, quotas, remaining budget, or cost forecasts.

### Customer Options for API Key Management

Enterprise customers can manage API consumption in two ways:

<table><thead><tr><th width="219.3333740234375">Option</th><th>Description</th></tr></thead><tbody><tr><td>OvalEdge-Provided API Key (Default)</td><td>Managed by OvalEdge; includes a $100 monthly quota. Additional usage is billed at $0.02 per call. Ideal for organizations seeking managed governance and consolidated billing.</td></tr><tr><td>Customer-Provided API Key</td><td>Customers generate and configure their own OpenAI key. Billing and quota management are handled directly by OpenAI. OvalEdge does not impose consumption restrictions in this case.</td></tr></tbody></table>

With either option, administrators can see the AI tokens actually consumed in the AI Consumption Dashboard. The dashboard works the same way whether the key is OvalEdge-provided or customer-provided.

### Security and Data Privacy in API Consumption

* All API keys are **encrypted** and securely stored; plaintext keys are never visible in OvalEdge systems.
* Each enterprise tenant is mapped to a **dedicated OpenAI project**, ensuring **customer isolation** and precise usage tracking.
* **Data confidentiality** is strictly maintained - OpenAI does not retain or train on customer prompts, ensuring privacy compliance.

## Public Edition Consumption and Usage

### Free Trial Account and Spend Limits

When a new user signs up for the **Public Edition**, a **free trial account** is automatically provisioned.

* Each trial account receives a **$10 spend limit** and remains valid for **14 days**.
* Once the trial expires or the spend limit is reached, further actions requiring compute or AI interaction are paused until the user upgrades to a paid plan.
* Higher plans, such as Business or Business Plus, include predefined limits - e.g., **$100** and **$500**, respectively.

This spend limit covers all forms of consumption within askEdgi, including:

* Workspace compute costs
* AI token consumption
* File uploads and associated **S3 storage charges**
* Recipe creation and execution charges

### Billing Triggers and Accumulation of Charges

Charges accumulate automatically based on the following triggers:

<table><thead><tr><th width="219.3333740234375">Trigger</th><th>Description</th></tr></thead><tbody><tr><td>Workspace Activation</td><td>Billing begins once the workspace is initiated (e.g., when the user opens the ask screen). Even if recipes are not yet executed, compute charges begin.</td></tr><tr><td>Recipe Execution</td><td>Each recipe execution deducts cost from the spend limit proportional to the compute and AI token consumption.</td></tr><tr><td>File Upload</td><td>Uploading any file initiates S3 storage costs that continue for the duration of the file’s presence in the workspace.</td></tr><tr><td>AI Interactions</td><td>Prompts, enrichments, or analytics tasks contribute to the AI token usage and are billed within the overall limit.</td></tr></tbody></table>

{% hint style="info" %}
Charges may continue to accumulate even during idle sessions if the workspace remains active. Users are advised to close inactive sessions to conserve spend limits.
{% endhint %}

### Recipe Consumption and Earnings

When a user executes a recipe:

* The corresponding execution cost is deducted from their credit balance.
* The recipe creator earns credits, creating a circular monetization ecosystem within the public Marketplace.

These transactions are reflected in the Usage and Earnings tabs under Manage Subscription, allowing users to monitor both consumption and income transparently.

## Monetisation (Public Edition)

The **Public Edition** uniquely supports creator monetisation through the **askEdgi Marketplace**.

* **Creators** can list recipes for free or paid consumption models (per-execution or subscription).
* **Consumers** spend from their credit limit when executing recipes, and a portion of that spend (typically **50%**) is credited to the creator.
* The system ensures **automated earnings tracking, credit settlement**, and **monthly disbursement** to creators.

This creates a self-sustaining ecosystem - encouraging community-driven innovation while maintaining transparent spend control.

## Billing Triggers and Charge Accumulation

The following events define when billing starts and how usage costs accumulate:

<table><thead><tr><th width="206.3333740234375">Component</th><th width="332.6666259765625">When Billing Starts</th><th>Applicable Model(s)</th></tr></thead><tbody><tr><td>Workspace Compute</td><td>When the user launches the workspace (askEdgi screen)</td><td>Public &#x26; Enterprise</td></tr><tr><td>Recipe Execution</td><td>When a recipe is triggered or run</td><td>Public &#x26; Enterprise</td></tr><tr><td>AI Token Consumption</td><td>For each AI-driven query, enrichment, or prompt</td><td>Public &#x26; Enterprise</td></tr><tr><td>File Upload (S3)</td><td>Upon uploading a file to the workspace</td><td>Public</td></tr><tr><td>API Usage (OpenAI)</td><td>When model inference occurs via integrated APIs</td><td>Enterprise</td></tr></tbody></table>

Each consumption type contributes to the monthly usage total and is reflected in the **Usage Tab** for both deployment variants.

### AI Enrichment Consumption

Every AI Enrichment execution, and every Validate Result run before it, is an AI call and adds to AI token consumption. AI Enrichment derives new columns from existing workspace data using an AI Function.

**What is sent to the AI model in one enrichment execution**

* The prompt written by the user.
* The values of the input columns the user selected.
* The output column names the user configured. One execution can populate several output columns, so a separate enrichment per output column is not needed.
* For the Prompt Analysis function only, any Additional Context the user selected: workspace objects and/or Data Stories. The complete Data Story content is used.

**How Additional Context affects consumption**

* Additional Context is sent to the model with the prompt and input columns, so it adds to the input tokens of each call.
* It is limited to 30% of the AI model’s available context window.
* Selected content is passed as-is; no deduplication is done between input columns, workspace objects, and Data Stories.
* The configuration screen displays a note advising users to include only small, relevant content, as it is shared with the AI.

**Validate Result**

* Validate Result is optional. It runs a sample through the same enrichment pipeline as execution, so it is also an AI call.
* Any change to output columns, input columns, prompt, or Additional Context clears the previous validation result. Execution always uses the latest configuration.

Existing single-output enrichments continue to run unchanged. Enrichment executions and Validate Result runs are recorded with their tokens and cost under **Audit Trails** > **AI** > **askEdgi**, Type **AI Functions**.

## Spend Limit Enforcement and Alerts

askEdgi ensures financial control through automated background jobs and usage monitoring:

<table><thead><tr><th width="220.666748046875">System Job / Service</th><th>Purpose</th></tr></thead><tbody><tr><td>Trial Reminder Job</td><td>Sends alerts before trial expiration (Public Edition).</td></tr><tr><td>Spend Limit Check Service</td><td>Monitors usage hourly and restricts transactions beyond defined thresholds.</td></tr><tr><td>Budget Threshold Alert (Enterprise)</td><td>Triggers when 80% of the quota is reached.</td></tr><tr><td>Budget Exhaustion Alert</td><td>Notifies users and admins when the budget limit is crossed.</td></tr><tr><td>Monthly Billing Service</td><td>Resets spend limits, populates billing tables, and refreshes quotas.</td></tr></tbody></table>

Alerts are also displayed contextually within the **askEdgi UI**, helping users act proactively before limits are reached.

## Usage Monitoring (My Usage)

The My Usage section provides transparency into current and historical consumption across major categories:

* **Compute** - Infrastructure usage and related costs.
* **AI Tokens** - Charges accumulated from AI-based queries and enrichments.
* **Recipes** - Recipe execution and associated costs.
* **Functions** - Charges for applied enrichments or transformations.

Hovering over graphs displays real-time charge details, allowing users to analyze month-on-month trends.

In Enterprise environments, usage reports can also be requested from the OvalEdge support team for audit or budgeting purposes.

Enterprise administrators can also review and export AI consumption records themselves under **Audit Trails > AI**. These views report AI tokens and per-call cost where it is recorded, alongside My Usage.

### Where AI Consumption Is Visible

All AI audit information is available in one place, under **Audit Trails > AI**. Each tab that covers more than one activity has a Type dropdown, so each view shows only the columns relevant to that activity.

<table><thead><tr><th width="168.9375">Tab</th><th width="286.484375">What it shows</th><th>Consumption detail</th></tr></thead><tbody><tr><td>AI Consumption Dashboard (default landing tab)</td><td>Consolidated token consumption, trend over time, and a log of individual AI operations</td><td>Input, output and total tokens</td></tr><tr><td>Embedding Usage</td><td>Embedding generation activity, with the source of each request</td><td>Input and output tokens per request</td></tr><tr><td>OvalEdge AI Activities</td><td>AI features outside askEdgi. Types: Description Generation (default), Data Classification</td><td>Data Classification records show input tokens, output tokens, and LLM Call Cost</td></tr><tr><td>askEdgi</td><td>All askEdgi activity. Types: User Prompts (default), AI Functions, Recipe AI Functions, Parameters Detection &#x26; Injection, Recipe Metadata Generation, Take Action</td><td>AI activity Types show input tokens, output tokens and EDGI LLM Call Cost per call. Take Action records are operational events with no token or cost fields</td></tr></tbody></table>

Changing the Type refreshes the grid with the matching records and columns without leaving the page. Search, filtering, sorting, export, and pagination work for every Type. The User Prompts view also shows running **Total Input Tokens** and **Total Output Tokens** for the records in view.

Access to these tabs is restricted. Users without the required Audit Trails permission cannot view AI audit information.

### AI Consumption Dashboard

The AI Consumption Dashboard shows how many AI tokens are consumed, which operation uses them, and how usage changes over time. All figures are calculated from the underlying AI audit records; nothing is estimated or projected.

**Summary metrics**

<table><thead><tr><th width="212.5390625">Metric</th><th>Definition</th></tr></thead><tbody><tr><td>Total Input Tokens</td><td>Sum of input tokens across the AI operations the dashboard covers</td></tr><tr><td>Total Output Tokens</td><td>Sum of output tokens across the AI operations the dashboard covers</td></tr><tr><td>Total Tokens Consumed</td><td>Total Input Tokens + Total Output Tokens</td></tr></tbody></table>

**Consumption by Operation Type:** A donut chart shows each operation’s share of total tokens consumed across askEdgi, MCP Query, Description Generation and Object Indexing. Legend entries are ordered from highest to lowest consumption and show percentage and token count. Hovering over a segment shows the operation type, total tokens and percentage contribution.

**Token Consumption Trend:** A time-series chart shows tokens consumed per period, viewable Daily, Weekly, or Monthly (default Monthly). It shows historical data only.

**AI Operations Audit Log:** A table lists the individual AI operations behind the dashboard figures, with pagination, sorting, search and filtering.

<table data-search="false"><thead><tr><th width="205.87109375">Column</th><th>Description</th></tr></thead><tbody><tr><td>Timestamp</td><td>When the AI operation ran</td></tr><tr><td>Operation Type</td><td>askEdgi, MCP Query, Description Generation or Object Indexing</td></tr><tr><td>User</td><td>User the operation is attributed to</td></tr><tr><td>Model</td><td>AI model used</td></tr><tr><td>Input</td><td>Tokens sent to the model</td></tr><tr><td>Output</td><td>Tokens returned by the model</td></tr><tr><td>Total</td><td>Input + Output</td></tr><tr><td>Status</td><td>Success, Failed, Pending or Partial Success</td></tr></tbody></table>

**Deployment behavior:** The dashboard behaves the same for Managed LLM, Bring Your Own Key (BYOK), SaaS, and On-Premise deployments because it reports only consumption that actually occurred.

**What the dashboard does not show:** Purchased or remaining tokens, consumption percentage, package expiry or renewal, forecasts, billing, quota management, or user rankings. Spend limits, quotas, and billing remain covered by My Usage and the alerts described above.

### Embedding Usage

The Embedding Usage tab shows embedding (vector) generation activity across the platform, with the source of each request.

<table><thead><tr><th width="205.453125">Source</th><th>Meaning</th></tr></thead><tbody><tr><td>Object Indexing</td><td>Embeddings for catalog objects such as tables, schemas, files, and glossary terms</td></tr><tr><td>Document Indexing</td><td>Embeddings for documents</td></tr><tr><td>Recipe Indexing</td><td>Embeddings for askEdgi recipes</td></tr><tr><td>MCP – Embedding</td><td>Embeddings generated for MCP requests</td></tr><tr><td>askEdgi – Discovery</td><td>Embeddings generated during askEdgi discovery</td></tr></tbody></table>

* Each record shows AI Engine, AI Model, Source (always populated), User Name, a reference to the object embedded (e.g., oetable:119936), AI Input Tokens, AI Output Tokens, Created Date, and LLM Call Details. An Error Message appears only if the call fails.
* **AI Input Tokens** are the tokens processed for the embedding. **AI Output Tokens** are 0, as embedding models produce no output tokens.
* The generated vector is not stored in the audit record, so Response from AI is not populated.
* The view reports tokens only; it has no cost column.

## Summary

**askEdgi’s** consumption and usage framework ensures transparency, control, and flexibility across deployment variants.

* **Public Edition** offers individual users an accessible entry point with defined spend limits and community monetisation options.
* **Enterprise Edition** provides a secure, quota-based consumption model with enterprise-grade governance, API isolation, and audit visibility, including a consolidated AI Consumption Dashboard under Audit Trails > AI.

This model ensures that every unit of resource consumed - from compute to AI tokens - is traceable, optimizable, and aligned with organizational or individual budgets.

***

Copyright © 2026, OvalEdge LLC, Peachtree Corners, GA USA


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.ovaledge.com/release8.3/askedgi/user-guides/consumption-and-usage-guide.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
