Colrows Pricing - Priced on Semantic Assets, Not Tokens or Seats

01 · How we price

Most tools bill you on things you can't forecast.

Tokens, queries, seats, CPUs - those are implementation details. They make your invoice volatile and your budget arguments hard to win. Colrows meters one thing: the size of the semantic layer we govern for you.

What we never meter

Volatile by design. Off your invoice by ours.

✗

Tokens consumed

LLM consumption swings with model choice and prompt shape. We absorb it - it never reaches your invoice.

✗

Seats provisioned

Every tier includes unlimited users. Adoption should be a win, not a line item.

✗

CPUs and environments

No vCPU counting. No environment counting. How you architect your deployment is your decision - and yours alone.

✗

Queries fired

Query volume is a fair-use band, not a billing meter. We protect our capacity; we don't monetise yours.

The one thing we price

Semantic Assets - the size of your governed graph.

✓

Semantic Assets

A Semantic Asset is any governed object maintained by Colrows within your semantic layer. That is the meter.

✓

Your tier

Determined by the total number of Semantic Assets in your Colrows graph. Nothing else.

✓

Coverage, not consumption

The conversation with us stays focused on what your graph covers - not on consumption accounting.

✓

One committed annual figure

You sign one number per year. It never changes without your approval.

02 · Tiers

Tiers scale with your graph. Not your headcount.

More Semantic Assets mean a larger semantic layer - that is what moves you up a tier. Every tier includes unlimited users, multiple data sources, and included fair-use query capacity.

Starter

Up to 500 Semantic Assets

A single team, one data source. The smallest governed graph.

Growth

Up to 2,000 Semantic Assets

Multiple teams, multiple sources. The graph broadens with your estate.

Scale

Up to 5,000 Semantic Assets

Cross-department deployment. One graph, governed end to end.

Enterprise

Custom asset count

The full organisation. Custom coverage, custom terms.

Every tier ships the full compiler: deterministic SQL generation, join path proof, compile-time governance, persona-scoped access enforcement, audit trails, and point-in-time reproducibility.

03 · Deployment

Two ways to run it. One pricing logic.

The Semantic Asset meter is the same in both. The only difference is where the compiler runs - and who runs the infrastructure.

Most teams start here

Shared SaaS

On our infrastructure

One figure fixed · annual

We manage the servers, the LLM, the updates, the monitoring. You connect your data sources and start querying in plain language. One fixed annual fee. Nothing more.

Sized by your Semantic Asset tier - Starter through Enterprise
Included fair-use query capacity - never a billing meter
Unlimited users, multiple data sources
Join path proof and governed query execution
Audit trails and point-in-time reproducibility
Never-exceed billing guarantee

Book a demo

On-Premise / Private VPC

Inside your environment

Three parts one predictable total

For strict data residency, compliance, or air-gap requirements. You run the infrastructure and the LLM. Colrows licenses the Semantic Compiler and supports your team. Data never leaves your environment.

Your Semantic Asset tier - same structure as SaaS
Self-hosted licence - private deployment rights, air-gap support, single-tenant architecture
Annual support - patches, upgrades, performance tuning
AWS VPC, Azure VNet, private GCP, or fully air-gapped
No vCPU counting. No environment counting.
Multi-year uplift locked at a hard 3 percent cap

Design your deployment

Built for BFSI, healthcare, pharmaceutical, government, and any regulated industry where data must stay on-premise or inside a private cloud.

04 · Commitments

Written into the contract. Not buried in the FAQ.

One committed number per year

You sign one annual figure. Your invoice never changes without your approval.

Never-exceed billing

We throttle at your cap. We never surprise-bill. Overage requires explicit opt-in from your team.

Model-independent pricing

Swap LLMs at any time without repricing. Your tier does not change if you change your model provider.

Token-price protection

We absorb LLM provider price changes. If your model provider raises prices, your bill stays the same.

A hard cap on renewals

Multi-year contracts cap annual uplift at 3 percent. No renewal shock.

90 days' notice

Pricing and tier definitions are reviewed annually. Any change is communicated to existing customers 90 days in advance.

05 · Proof of concept

Prove the value. Then commit.

Two POC paths. The goal of both is the same: demonstrate value against your actual data before you commit to anything.

Shared SaaS POC

Prove it on our cloud

Free for qualified prospects

Access to the Colrows Shared SaaS environment, with guided setup and a walkthrough session. Fair-use limits apply.

One data source connected
Semantic understanding and AI query execution, end to end
Guided setup and walkthrough session
Duration: 30 to 60 days

Start a POC

On-Premise / Private VPC POC

Prove it in your environment

Creditable in full toward year 1

A full deployment inside your infrastructure or private VPC, with dedicated support throughout. One-time engagement fee, credited in full toward your Year 1 licence on conversion.

Full deployment in your infrastructure or private VPC
One analytical data source onboarded
Full Consensus build - schema, relationships, enrichment
Knowledge-transfer and training sessions
Duration: 60 to 90 days

Scope your pilot

Questions? We have answers.

Don't find what you need here? Send us a message.

What is a Semantic Asset?

A Semantic Asset is any governed object maintained by Colrows within your semantic layer. Your tier is determined by the total number of Semantic Assets in your Colrows graph - nothing else. More assets mean a larger semantic layer, and that is what drives your tier.

Why don't you price on tokens, seats, or queries?

Tokens consumed, queries fired, seats provisioned, and CPUs spun up are implementation details. They make your invoice volatile and your budget arguments hard to win. Colrows prices on the one thing you can actually forecast and control: the size of the semantic layer we govern for you.

What happens if we run hot against the fair-use query band?

Every tier includes a query capacity. Queries slow progressively as you approach capacity to give you a clear signal, and pause at the hard cap. We never auto-bill overage. If you are running hot, we have a conversation about your tier - not a surprise line on your invoice.

Do you charge for users or infrastructure?

No. Every tier includes unlimited users and support for multiple data sources. On-premise, there is no vCPU counting and no environment counting - how you architect your deployment, how many tenants you create, and how many CPUs you allocate is your decision and yours alone.

What if our LLM provider raises prices?

We absorb LLM provider price changes - if your model provider raises prices, your bill stays the same. Pricing is also model-independent: swap LLMs at any time without repricing. Your tier does not change if you change your model provider.

Can we bring our own LLM?

Yes. Customers using self-hosted models, private Azure OpenAI deployments, or customer-managed Anthropic deployments may qualify for deployment-specific credits. Your platform licence stays unchanged - the Semantic Compiler, governed SQL generation, and orchestration runtime continue to run in full.

Do you support on-premise, private VPC, or air-gapped deployment?

Yes. For organisations with strict data residency, compliance, or air-gap requirements, Colrows deploys directly into your environment - AWS VPC, Azure VNet, private GCP, or fully air-gapped. You run the infrastructure and the LLM; Colrows licenses the Semantic Compiler and supports your team. Data never leaves your environment. Pricing combines your Semantic Asset tier with a self-hosted licence and annual support.

How do renewals work?

You sign one committed annual figure, and your invoice never changes without your approval. Multi-year contracts cap annual uplift at a hard 3 percent - no renewal shock. Pricing and tier definitions are reviewed annually, and any changes are communicated to existing customers 90 days in advance.

Can we run a proof of concept before committing?

Yes - two POC paths. The Shared SaaS POC is free for qualified prospects: one data source connected, semantic understanding and AI query execution end to end, over 30 to 60 days. The on-premise or private VPC POC runs 60 to 90 days inside your infrastructure with a full Consensus build and dedicated support, for a one-time engagement fee that is creditable in full toward your Year 1 licence on conversion.

What about customer-facing or embedded analytics?

Internal analytics usage is included in all licences - your employees, your analysts, your internal dashboards and BI portals. External-facing applications, commercial redistribution, and customer-facing experiences require an Embedded or OEM agreement, priced per engagement based on the application, audience, and distribution model.

One meter. One number. No surprises.

Colrows prices on Semantic Assets. Not the LLM. Not the servers. Not the CPUs. Not the seats. The number of objects we govern for you.

Book a demo Talk to sales

Priced on Semantic Assets. Nothing else.