Token Usage

NodeLoom tracks AI token consumption across your workspace with real-time quota enforcement and detailed monitoring analytics. This page covers how token tracking works, cost estimation, and how to configure limits.

How Tracking Works

Every AI node execution records the number of tokens consumed (prompt + completion). These totals are tracked per team, per month, and are used to enforce monthly quotas in real time. Usage data is also stored for the monitoring dashboard with per-execution, per-model, and per-workflow breakdowns.

Usage counters reset automatically on the first day of each calendar month.

Cost Estimation

NodeLoom estimates the cost of token usage based on published pricing for each model. Costs are calculated per 1,000 tokens and displayed in the monitoring dashboard:

Model	Cost per 1K Input Tokens	Cost per 1K Output Tokens
`gpt-4`	$0.03	$0.06
`gpt-4o`	$0.005	$0.015
`gpt-3.5-turbo`	$0.0005	$0.0015
`claude-3-opus`	$0.015	$0.075
`claude-3.5-sonnet`	$0.003	$0.015
`claude-3.5-haiku`	$0.001	$0.005
`gemini-1.5-pro`	$0.00125	$0.005
`gemini-1.5-flash`	$0.000075	$0.0003

Pricing accuracy

Cost estimates are approximate and based on published pricing at the time of each NodeLoom release. Actual costs depend on your provider agreement. You can override model pricing in the workspace AI settings.

Monthly Reset

Token counters reset automatically on the first day of each calendar month (UTC). Historical usage data is preserved indefinitely for reporting and trend analysis.

Limits and Warning Thresholds

Teams can configure monthly token limits and warning thresholds from the workspace settings:

Setting	Default	Description
Monthly token limit	Based on plan tier	Maximum tokens the team can consume per calendar month. Executions that would exceed the limit are rejected.
Warning threshold	80%	Percentage of the monthly limit at which a warning notification is sent to workspace admins.
Critical threshold	95%	Percentage of the monthly limit at which a critical notification is sent. Gives the team time to upgrade before hitting the hard limit.

Exceeding the limit

When a team exceeds their monthly token limit, all AI node executions are rejected with a clear error message. Non-AI workflow nodes continue to work normally. Workspace admins can temporarily increase the limit or upgrade the plan to restore AI functionality.

Dashboard Views

The token usage monitoring page provides three time-based views:

View	Period	Granularity
Daily	Last 30 days	Per-day token counts and estimated costs.
Weekly	Last 12 weeks	Per-week aggregated usage with week-over-week comparison.
Monthly	Last 12 months	Per-month totals with trend lines and budget tracking.

By-Model Breakdown

Each view includes a by-model breakdown showing the percentage and absolute token count per AI model. This helps teams identify which models are driving the most usage and cost. Use this data to:

Switch cost-sensitive workflows to smaller, cheaper models (e.g., GPT-3.5 Turbo or Claude 3.5 Haiku instead of GPT-4).
Identify workflows that consume disproportionately more tokens and optimise their prompts or tool usage.
Track the impact of model migrations on cost and quality.

Export

Token usage data can be exported for external analysis or billing reconciliation:

Format	Contents	Access
CSV	Date, workflow ID, workflow name, model, input tokens, output tokens, total tokens, estimated cost.	Download from the monitoring dashboard or via the API.
JSON	Same fields as CSV, structured as an array of objects.	Available via the API.

API export

Use the monitoring export endpoint with query parameters for date range and format. See the API reference for full details.

Next Steps

Anomaly Detection -- detect unusual token spikes on individual executions.
Drift Alerts -- catch gradual increases in token consumption over time.
Sentiment Tracking -- monitor conversational quality alongside token usage.

Sentiment

Authentication