Guides / culpa vs promptlayer
Culpa vs PromptLayer
PromptLayer is a prompt management and evaluation platform billed on requests, eval cells and agent node executions, with data retention control published only on Enterprise. Culpa, a local-first LLM cost, margin, and forecast ledger, prices every call from a versioned price book and forecasts next month from history you hold yourself.
Why this happens
PromptLayer meters more dimensions than most and publishes them clearly. Free is $0 with 5 users, 2.5k requests a month, 750 agent node executions and 250 eval cell executions. Pro is $49 a month with the same limits plus unlimited playgrounds and workspaces, and pay-as-you-go at $0.003 per transaction. Team is $500 a month with 25 users, 100k requests, 7.5k eval cells and $0.002 per transaction. The detail worth pulling out is on the feature table rather than the price cards: Data Retention Control appears under Enterprise and nowhere else, and self-hosting is Enterprise too. So on every published tier below Enterprise, how long your history lives is the vendor's decision rather than yours. For prompt work that's fine. For cost work it's the whole ballgame, because a forecast is made of history.
What this usually looks like
- You can't state your data retention because your plan doesn't expose it as a setting.
- Three separate meters move and only one of them is on your dashboard.
- A per-transaction rate applies to work nobody thinks of as a transaction.
- Prompt versions are well managed and nobody can price them.
- Self-hosting turns out to be a sales conversation rather than a setting.
Free, no card, no account
Run the free Cost Leak Scan
It shows your most expensive conversation before you install anything.
Mistakes that cost the most
| Mistake | Why it hurts | Do instead |
|---|---|---|
| Assuming retention is configurable on any paid plan. | PromptLayer lists Data Retention Control under Enterprise only, so the lower tiers inherit a default you didn't set. | Read the feature table rather than the price cards, and confirm retention in writing. |
| Budgeting from the request count alone. | Agent node executions and eval cell executions are separate published meters with their own limits. | Model all three meters against your own traffic before choosing a tier. |
| Treating prompt version history as cost history. | Knowing which version ran isn't knowing what it cost, and only one of those reconciles to an invoice. | Price each version's traffic from a rate book and compare on money. |
Run this check tonight
- Find your data retention in writing and check whether you can change it.
- List every meter your plan bills on and which ones you monitor.
- Price your two most-used prompt versions and compare them on cost, not score.
- Ask what your best prompt version costs per thousand calls.
What a per-transaction rate does at agent volume
PromptLayer publishes pay-as-you-go at $0.003 per transaction on Pro and $0.002 on Team. Agent workloads emit many logged steps per user request, which is the shape that makes per-transaction pricing bite. Take an agent averaging 8 logged steps per request. Rates are published. The step count and request volume are modelled.
Per-unit rates and platform fees cross over, and where they cross depends on a step count that nobody sets deliberately. Work out your own crossover before picking the tier with the lower headline rate, and price your model spend somewhere that doesn't move when you add a logging step.
Every number, with its confidence and source
| Figure | What it means | Confidence | Source |
|---|---|---|---|
| $500 per month | PromptLayer's published Team plan price, before per-transaction charges | provider-reported | promptlayer.com/pricing, read 2026-08-03. Team covers 25 users, 100k requests a month and 7.5k eval cell executions with pay-as-you-go at $0.002 per transaction. Pro is $49 a month at $0.003 per transaction. Free is $0. Data Retention Control and self-hosting appear on Enterprise only. |
| $1,200 per month | modelled per-transaction charges for an agent at 8 logged steps per request on the Pro rate | calculated | 50,000 requests x 8 logged steps = 400,000 transactions, x the published $0.003 Pro rate = $1,200. Request volume and steps per request are modelled. The per-transaction rates are published on promptlayer.com/pricing, read 2026-08-03. |
What a generic answer can’t know
PromptLayer does prompt management properly, with versioning, playgrounds, deployment approvals and evaluation, and none of that overlaps Culpa. The gap this page is about is retention and ownership. PromptLayer publishes Data Retention Control and self-hosted deployment on Enterprise only, so below that tier the length of your own history is decided for you. Culpa keeps the ledger as a database on your infrastructure by default, so how far back you can look is a disk decision at every size. On top of that it prices each call from a versioned price book in exact decimal, holds the revenue each customer pays you so spend becomes margin, and forecasts next month with a range. Manage prompts there. Answer the money question here.
Questions founders ask next
What does PromptLayer cost?
Free is $0 with 5 users, 2.5k requests, 750 agent node executions and 250 eval cell executions a month. Pro is $49 a month with pay-as-you-go at $0.003 per transaction. Team is $500 a month with 25 users, 100k requests and $0.002 per transaction. Enterprise is custom. Read 2026-08-03.
Can I control data retention on PromptLayer?
Its published feature table lists Data Retention Control under Enterprise and not on Free, Pro or Team. Self-hosted and managed single-tenant hosting are also Enterprise. On the lower tiers, retention is a default you inherit rather than a setting you choose.
Does PromptLayer track LLM cost?
Its published tiers meter requests, agent node executions and eval cell executions rather than publishing a model rate book, and nothing on its site claims margin against revenue or a forecast of spend, as read 2026-08-03. Treat it as prompt and evaluation tooling.
Do PromptLayer and Culpa compete?
Not really. PromptLayer manages and evaluates prompts. Culpa prices the traffic those prompts generate, attributes it to customers against revenue, and forecasts the month. The overlap is that both see every call, and they do entirely different things with it.
On your infrastructure
Culpa runs on your infrastructure. Your prompts and responses never leave it. Culpa counts calls to run your plan, and it fails open, so if it ever breaks your app keeps running.
How Culpa works
Find the culprit. Not just the total.
Your dashboard shows what you spent. It stops short of who spent it. Culpa shows the conversation, the user and the feature behind it.
Your prompts stay local.
Culpa runs on your own infrastructure. What you send to a model reaches us at no point.
Every dollar has a name.
Follow any charge to the conversation, the user, the feature and the customer behind it.
See the bill before it lands.
Cost your next feature before you ship it. You get the likely bill and the worst case, at best, median, p90 and p99.
Three steps to your first answer.
Change one base URL.
Or drop in the Python or TypeScript library.
Find your most expensive conversation.
In the first session, not the first week.
Cost your next feature before you ship it.
Why the bill went up
Example dashboardCalls traced
418,209
across 3 projects
Spend this week
$378.41
+ $182 vs last week
Failed calls
312
74% retried, and you paid for all of them
+ $182 this week traced to one culprit
Spend over 14 days
Most expensive users
Next week forecast
Graded against reality. Accuracy shown as results land.
Free, no card, no account
Run the free Cost Leak Scan
It shows your most expensive conversation before you install anything.
Keep reading
Sources: PromptLayer, PromptLayer pricing. Last reviewed 2026-08-03. Plain text version.