Per-run agent platforms

Krazimo vs Lyzr

The same unit, the same honesty about inference, and a lower price. The difference worth arguing about is what the run leaves behind.

The same billing unit

One agent turn, so the two price lists can be compared directly instead of translated.

Lower at the published rate

$99 a month covers 5,000 runs.

Retrieval that cannot fabricate

Navigation over an index, not similarity over embeddings.

We copied the unit on purpose

Lyzr publishes a per-agent-run price of $0.081 on their managed cloud, and states that model tokens are billed separately. We do both of those things, for the same reason: a run is the unit a buyer can forecast, and a vendor taking a margin on inference is a toll on your growth. Where a competitor has got the commercial model right, the useful thing to do is say so and compete on the part that is actually different.

The comparison is cloud to cloud

Lyzr also publishes $0.032 for VPC and on-premises deployments, and labels it their best value. We compare against the cloud rate because that is like-for-like: on VPC the customer brings and pays for the compute, and we do not sell a self-hosted plan today. If you are weighing a self-hosted deployment, their VPC rate is the number to use, and you should hold that against the fact that the engine underneath our product is a self-contained container we intend to publish under Apache-2.0.

What differs is inside the run

A per-run price says nothing about how the run reaches its answer. Ours navigates a document tree and opens pages by name, so a path no listing offered is refused by the tool rather than discouraged by an instruction, and a citation lands on the document rather than on a chunk with a similarity score. Then deterministic guards run on the finished answer, each reported by name, and one version number pins the library, the policy, the evaluation suite and the model together.

The arithmetic

The arithmetic, at their published rate

At 5,000 runs a month, $0.081 per run is $400 in platform fees against our $99. At 40,000 it is $3,200 against $499. Platform fee only on both sides — both of us bill inference separately, so it cancels out of the comparison rather than being hidden by it.

Their VPC rate would put the same volumes at $150 and $1,200, against a compute bill you pay yourself. We think cloud-to-cloud is the honest line; you may not, so the number is here either way.

Depending on who is asking

Platform engineering

“How much of this do we operate?”

No vector database to run, size or re-index. Documents go in a tree; nothing is chunked or embedded before it can be used.

Compliance

“Can I sign off on what it said?”

One version pins the documents, the policy, the suite and the model, records who published it, and can be restored. The evidence trail is the product.

Finance

“Will the bill be forecastable?”

A flat run, however many steps it takes, and no margin on model spend. Usage is returned on the call that incurred it, so the invoice reconciles against your own logs.

What Lyzr publishes

Every line here is a figure from their own page, with the wording it was read from in the footnotes. Nothing in this table is our characterisation of their product.

	Per-run agent platforms
Billing unit	Per agent run1
Cloud rate	$0.08 per agent run1
VPC / on-premises rate	$0.03 per agent run2

Platform fee at 5,000 agent runs a month

Their published cloud rate against our plan, on the same unit. Platform fee only — both of us bill inference separately, so it cancels out rather than being hidden.

LyzrKrazimo Studio

How Krazimo Studio is built

Our own architecture, stated as such. Whether it is better than the alternative is your call to make; these are the facts you would be deciding on.

Retrieval navigates a named index — no embedding step, and no vector store to run or re-index
A path no listing offered is refused by the tool itself, before any text is generated
Guards are deterministic rules on the finished answer, each reported by name with its reason
One version pins the library, the policy, the evaluation suite and the model, and can be restored
Usage is returned on the call that incurred it — the same number that bills
Unlimited agents on every plan, and no share of your model spend

Questions we get

Is a run the same thing on both sides?

Close enough to compare. Ours is one agent turn — one question in, one governed answer out, with every step it takes inside that single run. Lyzr publishes per agent run as well, which is why this is the one comparison on the site that is arithmetic rather than argument.

Do either of you mark up inference?

Neither. Lyzr states tokens are billed separately at market rates; ours are passed through at cost or billed directly to your own provider key.

Why not compare against their VPC price?

Because it excludes the compute you would be paying for, and we do not sell a self-hosted plan to compare it against. The number is on this page so you can make that judgement yourself.

Sources

Every external figure on this page, with the wording it was read from and the date it was read. Competitor prices change; if one of these is out of date, tell us and it gets corrected here rather than argued about.

1Lyzr, pricing page
“$0.08/agent run”
lyzr.ai/pricingread 2026-09-07
2Lyzr, pricing page
“$0.03/agent run”
lyzr.ai/pricingread 2026-09-07

Other comparisons

Krazimo vs Amazon Bedrock Krazimo vs Microsoft Foundry Krazimo vs Salesforce Agentforce

Every number on this page has a source. That is the product, in a different medium.

$99 a month to start. Unlimited agents, no share of your model spend, and a record a compliance owner can sign.

Open the console See the price list