Usage & Analytics

Monitor resource consumption in real time, analyze usage events, understand token mechanics, and track costs across all your companies

Overview Dashboard

Navigate to Billing → Overview to open the Usage Analytics dashboard — a single page that gives you full visibility into resource consumption, platform limits, and a detailed log of every billable event.

The dashboard is divided into four sections: Plan Summary, Resource Usage, Platform Limits, and Usage Events.

---

Plan Summary

The top of the page displays a hero card with your current subscription details.

Left side:

• A gradient badge showing your plan name (Free, Premium, or Professional) with a crown icon
• If on a paid plan — the monthly price (e.g., $29 /mo) and the next payment date (e.g., “Next payment: 15.04.2026”)
• If on the Free plan — a prompt to upgrade

Right side:

• Manage button — opens the Manage Subscription modal (paid plans only). See Managing Your Subscription
• Upgrade Plan / Change Plan button — links to the Pricing page

---

Resource Usage

Six progress-bar cards arranged in a 3-column grid (2 columns on tablet, 1 on mobile). Each card tracks one billable resource in real time.

#	Resource	Icon	What It Measures
1	Standard Tokens	Bolt (blue)	Tokens consumed by standard models — GPT-4o Mini, Claude 3.5 Haiku, Gemini 2.0 Flash, Llama 3.1 8B
2	Premium Tokens	Bolt (purple)	Tokens consumed by premium models — GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Pro, Llama 3.3 70B
3	Voice STT	Microphone (cyan)	Speech-to-Text transcription time in minutes
4	Voice TTS	Speaker (amber)	Text-to-Speech synthesis in characters
5	Embeddings	Database (green)	Tokens used for knowledge base training (chunking + vector generation)
6	Storage	Folder (rose)	Total file storage across all your companies

Each card displays:

1. Header row — the resource icon and label on the left, a percentage badge on the right
2. Progress bar — an 8px-tall bar showing current consumption as a filled gradient
3. Values row — “N used” on the left, “of N” (your plan limit) on the right

Color Thresholds

The percentage badge and progress bar change color based on how much you’ve consumed:

Usage Level	Color	Meaning
Below 80%	Green	Normal — you have plenty of headroom
80% – 89%	Amber	Warning — consider upgrading or buying a pack
90% and above	Red	Critical — approaching the limit, requests may be blocked soon

Storage values are displayed in human-readable units — MB for amounts under 1 GB, GB for larger values. All other resources show formatted numbers with thousands separators.

---

Platform Limits

Four counter cards below the resource meters, arranged in a 4-column grid (2 on tablet, 1 on mobile). These show how many platform resources you’ve created versus your plan maximum:

Icon	Counter	Example
Building (blue)	Companies	2 / 3
Wand (purple)	AI Agents	4 / 5
Document (green)	Knowledge Documents	45 / 100
Messages (amber)	Conversations	1,200 / 5,000

When a limit is unlimited (e.g., AI Agents on the Professional plan), the maximum displays as ∞.

When you attempt to create a resource that exceeds your plan limit (e.g., creating a 6th agent on the Premium plan), a modal appears explaining the limit and offering to upgrade or purchase a pack.

---

Usage Events

The bottom section contains a detailed, filterable table logging every billable action across all your companies. This is your audit trail for understanding exactly what consumed your quota.

Filtering

Click the + Filter button to add filter chips. Six filter types are available:

Filter	Type	Options
Date From	Date picker	Calendar popover — select a start date
Date To	Date picker	Calendar popover — select an end date
Company	Select dropdown	Lists all your companies by name
Event Type	Select dropdown	Chat, Workflow, STT, TTS, Embedding, Storage
Resource	Select dropdown	Standard Tokens, Premium Tokens, STT Minutes, TTS Characters, Embedding Tokens, Storage
Channel	Select dropdown	Admin, Widget, System

Active filters appear as removable chips above the table. Removing a chip immediately reloads the table with updated results.

Table Columns

Each row in the table represents a single billable event with eight columns:

Column	Description	Format
Date	When the event occurred	DD.MM.YYYY HH:mm
Company	Which company’s agent triggered the event	Company name, or — if system-level
Event	Type of billable action	Color-coded badge (see below)
Resource	Which resource quota was consumed	Plain text label
Channel	Where the interaction originated	Grey badge (Admin, Widget, System), or —
Model	Which LLM model was used	Monospace font (e.g., gpt-4o-mini), or — for non-LLM events
Amount	Quantity consumed	Smart-formatted (see below)
Cost	Estimated cost in USD	Bold, tabular numerics (see below)

Event Type Badges

Each event type has a distinct color-coded badge for quick visual scanning:

Event Type	Badge Color	Description
Chat	Blue	LLM inference from a chat conversation
Workflow	Purple	LLM inference from a workflow node execution
STT	Cyan	Speech-to-Text transcription
TTS	Amber	Text-to-Speech synthesis
Embedding	Green	Knowledge base training (vector generation)
Storage	Rose	File upload or storage operation

Amount Formatting

The Amount column adapts its display format based on the resource type:

Resource	Format	Example
Tokens (Standard, Premium, Embedding)	Comma-separated number + “tokens”	4,521 tokens
STT Minutes	Minutes and seconds	2m 34s
TTS Characters	Comma-separated number + “chars”	1,250 chars
Storage	MB or KB	3.2 MB, 450 KB

Cost Formatting

The Cost column shows estimated USD cost with precision adapted to the value:

• **$0.01 and above** — displayed with 2 decimal places (e.g., `$1.25`)
• **Below $0.01** — displayed with up to 8 decimal places, trailing zeros removed (e.g., `$0.00042`)
• Zero or not applicable — displayed as —

Pagination

The table loads 25 events per page and uses asynchronous pagination. The bottom bar shows:

• “Showing X–Y of Z” — current range and total count
• Page buttons — first page, last page, and a window of ±2 pages around the current one, with ellipsis (…) for gaps

While loading, the table displays an animated skeleton with 8 shimmer rows. If no events match your filters, an empty state message appears: “No usage events found.”

---

How Tokens Work

Every AI interaction consists of two parts:

• Input tokens — your message + system prompt + knowledge base context (RAG chunks) + conversation history + tool descriptions
• Output tokens — the agent’s response

Both count toward your quota. The total per interaction depends on:

1. System prompt length — longer prompts consume more input tokens per message
2. Knowledge base context — more RAG chunks retrieved = more input tokens. Control this via RAG Settings → Chunk Limit
3. Conversation history — longer conversations accumulate context from previous messages
4. Response length — controlled by the Response Length personality slider in agent settings
5. Tool calls — each tool call round adds input tokens (tool descriptions) and output tokens (tool arguments + results)

Token Types

Token Type	Models	Cost Profile
Standard	GPT-4o Mini, Claude 3.5 Haiku, Gemini 2.0 Flash, Llama 3.1 8B	Lower cost, faster responses
Premium	GPT-4o, Claude 3.5 Sonnet, Gemini 2.5 Pro, Llama 3.3 70B	Higher cost, better quality
Embedding	text-embedding-ada-002	Used only during knowledge base training

To optimize token usage: keep system prompts concise, limit the number of RAG chunks per query (RAG Settings → Chunk Limit), and use standard models for simple tasks while reserving premium models for complex responses.

---

Voice Usage

Voice interactions consume two separate resources:

• STT (Speech-to-Text) — measured in minutes. Each voice message is transcribed, and the duration (rounded up to the nearest minute) is deducted from your quota.
• TTS (Text-to-Speech) — measured in characters. The agent’s text response is synthesized to audio, and the character count is deducted.

A single voice conversation turn typically consumes both STT minutes (the visitor’s audio) and TTS characters (the agent’s spoken response). These appear as separate rows in the Usage Events table — an STT event (cyan badge) and a TTS event (amber badge).

---

Storage Usage

Storage is consumed when you upload files to the File Manager:

• Documents — PDF, Word, Excel, TXT
• Photos — JPG, PNG
• Videos — MP4, WebM, MOV
• Audio recordings — conversation audio files

Storage usage is tracked bidirectionally — uploading increases usage, deleting files decreases it. The Storage meter on the dashboard shows your current total across all companies.

---

Quota Notifications

Revol monitors your usage and sends alerts before you hit limits:

Threshold	Level	Notification
80%	Warning	Email + in-app banner: “You’ve used 80% of your [resource]. Consider upgrading or purchasing a pack.”
90%	Critical	Email + in-app banner: “You’ve used 90% of your [resource]. Requests will be blocked at 100%.”

Each alert fires once per billing period per resource type. Critical alerts take priority — if your usage jumps from 75% to 92%, only the critical notification is sent.

In-app banners appear at the top of every page in the dashboard. You can dismiss them for the current session, but they reappear on your next login if the threshold is still exceeded.

---

What Happens at the Limit

Resource	Behavior
Standard / Premium Tokens	Agent stops responding until the next billing cycle or until you purchase a pack
Storage	Cannot upload new files to the File Manager
Voice (STT/TTS)	Voice features are disabled, text chat continues working
Conversations	New conversations are blocked, existing ones continue
Knowledge Docs	Cannot add new documents to agent knowledge bases

When a limit is reached and you attempt to perform a blocked action, a modal appears showing your current usage, plan limit, and options to either upgrade your plan or purchase an add-on pack.

---

Billing Period

• Free plan — daily billing period. Token usage resets every day at midnight UTC.
• Paid plans — monthly billing period. Usage resets on the anniversary of your subscription date (e.g., if you subscribed on March 15, your next reset is April 15).

Add-on packs are independent of the billing period — they have their own 3-month expiration from the purchase date. When your included plan quota is exhausted, consumption automatically switches to your oldest-expiring pack.