Your provider dashboard shows one number: total spend. TokenFlowIQ shows the gross margin on each AI feature, which customers are burning more than they pay, and what your waste is costing you — in dollars, from your real traffic.
No credit card. Free tier stays free.
Every LLM provider will tell you what you spent. None of them will tell you whether you made money on it.
Attribute every token to the feature that spent it. When the bill moves, you know what moved it — on the day it moves, not at the end of the month.
Per-customer AI cost against what they actually pay you. Catch the ones eating their subscription before renewal, not after.
Margin by feature and by customer, reconciled against the provider's own reported costs — so the number survives the follow-up question.
Retries, oversized models, cache misses, TTL mistakes — ranked by what each one costs you per month, measured from your traffic rather than assumed.
No proxy in front of your model calls, and no rewrite of how you call them.
Drop in the Python SDK, or import history from your provider's export. Events carry the feature and customer they belong to.
Tell us what your own plans charge. That's the half nobody else has — and it's what turns spend into margin.
Margin by feature and customer, waste ranked by dollars, and alerts when a budget or a margin floor is crossed.
# pip install llmcost-sdk
import llmcost_sdk as tfq
tfq.init(api_key="llmc_ik_...")
client = tfq.wrap_anthropic(client) # or wrap_openai
# Cost lands on the feature and the customer that caused it
with tfq.tag(feature="doc-summary", customer_ref="acme-corp"):
resp = client.messages.create(model="claude-opus-4-5", ...)Tiers are named for team size because that's how teams describe themselves, but tracked LLM spend is the meter — whichever limit you hit first moves you up. No seat charges inside a tier: a five-person plan costs the same whether one person logs in or five.
| Tier | For | Seats | Tracked spend / mo | Retention | Monthly | Annual |
|---|---|---|---|---|---|---|
| Free | Evaluating, side projects | 1 | ≤ $2,000 | 30 d | $0 | — |
| Solo | An individual shipping AI features | 2 | ≤ $10,000 | 90 d | $49 | $490 |
| Team | Small teams, up to 5 people | 5 | ≤ $50,000 | 12 mo | $249 | $2,490 |
| Scale | Medium teams, up to 25 people | 25 | ≤ $250,000 | 24 mo | $999 | $9,990 |
| Enterprise | 50+ developers, or > $250k/mo | Unlimited | Unmetered | 24 mo+ | Talk to us | Annual commit |
Annual is two months free. No credit card on Free.
TokenFlowIQ is in private beta. If your LLM bill is growing faster than you can explain it, we'd like to talk — beta accounts are hand-held and shape what ships next.
hello@tokenflowiq.comTell us roughly what you spend per month and what you're building. We answer every email.