The AI tab in the monitoring dashboard provides a clear overview of AI usage across your account. It helps you track API calls, estimate costs, and monitor spending against your configured limits.
The top of the dashboard summarizes your AI usage over the last 30 days:
Estimated spend: The total dollar cost across all AI model requests.
Tokens: The total number of tokens consumed (input + output).
Turns: The total number of individual requests sent to AI models.
Cache hit rate: The percentage of input tokens served from the cache, which helps reduce costs.
Key benefits
Monitor costs: Identify which models and teams drive your AI spending.
Enforce budgets: Compare usage against your set limits.
Optimize usage: Pinpoint expensive models for potential optimization.
Analyze caching: Track how much you are saving through prompt caching.
Compare providers: Easily switch between providers to compare performance and costs.
Filter by providers
Use the dropdown menu to analyze specific AI providers:
All providers: Aggregates usage across every configured provider.
Managed by the platform: AI providers managed automatically by our system.
Custom providers: Any AI providers you have added manually in your AI settings.
Spend vs. limits
This visual donut chart compares your actual AI spend against your configured budget limits.
Total spend: Displays the total dollar amount for the last 30 days.
Visual breakdown: Shows segments for team spaces, personal spaces, and unattributed spend.
Budget tracking: Monitors both provider-level and space-level spending caps.
Spend by model
This bar chart breaks down your costs by specific AI model, sorted by highest cost first. It tracks usage across workflow builds, workflow runs, chats, and other activities.
Usage breakdown
There are three views available within this panel. Use the toggle at the top of the panel to switch between three views:
Models. Tokens, cache hit rate, and estimated cost per model.
Workflows. Your top workflows by AI usage, with cost split by activity.
Spaces. Estimated cost per space, so you can see which teams drive spending.
Models
This table provides a detailed view of your metrics per model.
Model: The specific AI model used.
Turns: Total number of requests.
Tokens: Total token consumption.
Cache: Your cache hit rate percentage.
Est. cost: The estimated dollar cost for that model.
Workflows and spaces
The workflows and spaces views break each estimated cost into where the usage happened:
Build. AI usage from building workflows.
Run. AI usage from workflow runs.
Monitor. Usage from autofix and autotune.
Other. Usage outside a workflow's lifecycle, such as personal chat.
Est. cost. The total across all four activities.
Workflows view
This view ranks your top 100 workflows by AI usage for the selected time window, across all providers. Each row shows the workflow's name, the space it belongs to, and its estimated cost split across build, run, monitor, and other.
Select a workflow to open it, then use the health and execution details in the editor to dig into what is driving its cost.
Note: Workflows deleted during the window still appear so their spend is not lost, listed as "Deleted workflow."
Spaces view
This view totals estimated AI cost per space, so you can see which teams account for the most spend. Personal spaces are listed by their owner, which helps you attribute individual usage. Alongside your spaces, you may see two summary rows:
Personal chat. AI usage from personal chat, which has no workflow and counts as Other.
Unattributed. Usage that is not tied to a space and is not personal.
Select any space to open it and review the workflows inside.
How costs are calculated
The estimated spend shown here is calculated by Tines 3B from the token usage it meters, multiplied by the pricing you configure.
Note: It is an estimate based on your rates, not an amount billed by the model provider. Tines 3B does not pull figures from a provider's billing or cost API, so the numbers you see are only as accurate as the rates you set.
For each model on a connector, 3B applies a separate rate to input tokens, output tokens, cache read tokens, and cache write tokens, then adds them up. Caching is accounted for: because the input token total already includes cached tokens, only the fresh (uncached) portion is charged the input rate, while cache reads and writes are billed at their own rates. Usage is tracked at a daily granularity.
You control these rates in two ways:
Set your own cost per token: in a connector's measure pricing section, anyone with permission to manage AI providers can enter a rate for each measure the connector meters. Token rates are entered per 1 million tokens, requests per 1,000, and data transfer per GB. These rates drive both the estimated spend shown here and usage against your spend limits.
Rely on editable defaults for known models: for recognized models, Tines 3B pre-fills list pricing as a placeholder. An unedited field still prices using that default, and typing a value overrides it. For Explore edition users, platform-managed providers have no rates to enter and are priced from published list-price estimates.
Note: If a model has no configured rate and no default, its cost cannot be estimated and the dashboard displays a placeholder instead of an amount. To fix a blank or inaccurate figure, add or correct the rate in that connector's measure pricing.
Data timeframe
All data in the AI tab defaults to the last 30 days. This can be changed in the drop down menu.
