ObservabilityComing soon

Dynamic tags

Attach labels to each request so you can slice usage and cost by feature, customer or team.

Coming soonLabels per requestSlice usage and cost

What is Dynamic tags?

Dynamic tags are labels you attach to each AI request at runtime, such as feature, customer, environment or team, so usage and spend can be split by those labels. Spro's dynamic tags are planned and not yet available. Until they ship, trace names and API keys are the closest way to group spend.

A single monthly total does not say which feature or customer caused the cost. Dynamic tags put that context on each request, so spend can be sliced the way your business is organised.

This page describes planned behaviour. Tag limits and how tags appear in reports will be published with the feature.

Key features

What this feature does for your requests.

  • Attach labels to each request at runtime
  • Slice usage and cost by tag
  • Labels such as feature, customer, environment or team
  • Tags shown in usage and spend views

How it works

4 steps, from setup to the first request.

  1. 01

    Choose your labels

    Decide which dimensions matter, such as feature or customer.

  2. 02

    Send tags with requests

    Add the labels to each call.

  3. 03

    Filter the usage view

    Group spend by any tag.

  4. 04

    Act on the numbers

    Find the feature or customer driving cost.

Is it the right fit?

Specific situations where this feature helps, and where it does not.

Good fit if

  • You bill AI cost to customers, features or teams
  • Finance needs spend split by business unit
  • You want to compare the spend of two product features

Not a good fit if

  • One product with one budget owner
  • You can already split spend by API key or workspace
  • Your team will not send tags consistently, so the data would be incomplete

Dynamic tags FAQs

Common questions about Dynamic tags.

Route every AI request through SPRO.

One OpenAI-compatible API for 350+ models, with intelligent routing, 29 built-in guardrails, observability, and automatic failover.

Talk to sales