Comparison
Flitch vs Looker
Google Cloud’s model-first BI platform against an AI-first builder for dashboards, reports and apps. What LookML buys you, and what it costs to keep.
Which should you choose?
Flitch has a semantic layer you set up without code, a published price, and a dashboard the same afternoon. Looker’s equivalent is LookML, which a developer writes and keeps writing, and Google publishes no price at all. Looker is the better choice when you have that team and need one governed model across a large company.
Checked against both products on .
Flitch vs Looker at a glance
| Flitch | Looker | |
|---|---|---|
| Published price | Yes: free, $20/user, or $40/user | No. Contact sales for a quote |
| Pricing shape | Per user, monthly with no contract, or 20% off annually | Platform fee plus per-user seats, on 1, 2 or 3 year terms |
| Editions | Free, Pro, Business, Enterprise | Standard (up to 50 internal users), Enterprise and Embed |
| Seats included | Priced per person who builds; viewers and published-dashboard readers need none | 10 Standard and 2 Developer users per instance |
| How the numbers are computed | Generated SQL, aggregated at the source against the semantic layer | SQL generated from LookML, aggregated in your warehouse |
| Costs beyond the sticker price | None. Credits are the whole bill | Your warehouse compute, since Looker queries it, plus AI data tokens when the promotional period ends |
| Seat types | Admin, member and viewer; only the people who build are charged for | Three: Developer, Standard, Viewer, each with different access |
| Free tier | Yes | No |
| AI included | Generation is the product. The plan fee is the credit allowance: $20 buys 200 credits worth $20 | Conversational Analytics, free while Google extends its promotional period, then a monthly data-token quota with $3/1M input and $20/1M output above it |
| Integrations | 61, including file upload, an AI API finder and input tables | Warehouses, via the model |
| Semantic layer | Yes, relationships and calculations, no code | LookML, written and maintained as code |
| Who sets it up | Anyone, in plain language | A developer, in LookML |
| Time to first dashboard | Minutes | Weeks, once the model is built |
| Warehouse required | No | Yes, and it is happiest on BigQuery |
| Public publishing | Custom domains, password protection, email-restricted sharing | Embedding and scheduled delivery, seat-bound sharing |
| Governance | SSO and an exportable audit log on Business; row security on the dataset, one rule per dataset; SCIM only under an Enterprise agreement | Mature, including row-level access in the model |
| Acting on the data | Workflows included: triggered steps that transform rows, then email them, post them to a webhook or save them as a dataset | Scheduled delivery of a Look to email, webhook, S3 or SFTP, threshold alerts, and other services through the Action Hub |
What does Looker cost?
Google publishes no dollar figure at all. The pricing page sets out the shape, a platform fee to run the instance plus per-user seats across Developer, Standard and Viewer roles, in Standard, Enterprise and Embed editions, each including 10 Standard and 2 Developer users, on annual subscriptions of one, two or three years. Then every price cell says to call sales. That is normal for enterprise BI and it is also a real difference: you cannot find out what Looker costs from its website, and you are committing for at least a year when you do. Flitch is free, $20 per user, or $40 per user on Business, monthly with no contract, or 20 per cent less if you choose to pay for a year. One of those answers is available this afternoon. Google does publish AI pricing, though, and it is the one part you can read: Conversational Analytics sits in a promotional period with no quota or overage fees, which Google has extended and will end with 90 days of notice, after which a Standard subscription gets 60M input and 1.2M output data tokens a month and pays $3.00 per million input and $20.00 per million output above that.
Is LookML worth it?
If you have the people, yes, and it deserves credit rather than a dismissal. LookML is a mature language for defining data models that fixes your metrics in version-controlled code, so every dashboard and every ad-hoc question resolves to the same definitions. For a company with real analytical complexity and a data team to own it, that is genuinely valuable and it is why Looker is where it is. The cost is that somebody has to write it and keep writing it. LookML is a project, not a setup step, and if it goes unmaintained the model drifts from the business and the dashboards quietly stop being right. Flitch has a semantic layer with relationships and calculations that you define without code, which is less expressive and needs nobody to own it.
What do you actually do in each one?
In Looker, a developer defines the model, and then business users explore within it, building Looks and dashboards from governed dimensions and measures. The exploration is good and the guardrails are the point. In Flitch you describe the dashboard you want and it is generated, then you refine it in plain language or edit the code directly. Looker gives a large company a safe, consistent way for many people to ask questions. Flitch gives a small one a dashboard today. Those are different problems, and picking the wrong one is expensive in opposite directions.
Which suits a small team better?
Flitch, fairly clearly. Looker Standard is aimed at under fifty users but still carries a platform fee, a sales process and a data model requirement, and none of those scale down well to five people. A small team usually does not have a LookML developer and will not hire one for a dashboard. Flitch has a free tier, published prices and no data model to build first, which is the practical difference between starting this afternoon and starting next quarter.
Does it matter which warehouse you use?
More than it should. Looker works across warehouses, but it is a Google Cloud product and the experience, integration and commercial packaging are strongest on BigQuery. If you are already on Google Cloud that is an advantage worth having. If you are not, some of Looker’s appeal is being paid for and not used. Flitch connects to 63 sources directly, including databases, warehouses, files and mainstream SaaS applications, and does not need a warehouse to exist at all.
Choose Looker if
- You have data engineers who will write and maintain LookML
- You need one governed definition of every metric across a large company
- You are committed to Google Cloud and want BI packaged with it
- You have many people asking ad-hoc questions who need guardrails
- You need access rules expressed in the model, not one filter per dataset
- Enterprise procurement and a sales process are normal for you
Choose Flitch if
- You do not have anyone to own a data model
- You want to know the price without booking a call
- You want a semantic layer with relationships and calculations without writing code
- You need a dashboard this week, not after a data model project
- Your data is not in a warehouse, or not in BigQuery
- The dashboard is going to clients or onto your own domain
- You would rather the AI ran on a provider account you control: Enterprise agreements support your own key, billed to you at your rates
Flitch vs Looker FAQ
How much does Looker cost?
Google publishes no figure. Pricing combines a platform fee with per-user seats across Developer, Standard and Viewer roles, across Standard, Enterprise and Embed editions, each including 10 Standard and 2 Developer users, on one, two or three year terms, and every price requires a call with sales. Flitch publishes its prices: free, $20 per user per month, or $40 per user on Business, with no contract.
Does Looker charge extra for AI?
Not yet. Conversational Analytics runs in a promotional period with no quota or overage fees, which Google has extended and says it will end only after 90 days of notice. Then each subscription gets a monthly data-token quota, 60M input and 1.2M output on Standard, with $3.00 per million input and $20.00 per million output above it. In Flitch the plan fee is the AI allowance: $20 buys 200 credits worth $20.
Do I need to know LookML to use Flitch?
No. Flitch has no language for defining data models. You set up relationships and calculations without code, and describe dashboards in plain language. That is less expressive than LookML for a large company and considerably faster for a small one.
Is Flitch a Looker replacement for an enterprise?
No. Flitch has SSO, an exportable audit log on Business, and row security that filters each viewer to the rows matching them. What it does not have is SCIM outside an Enterprise agreement, and its rule is one filter per dataset rather than access grants expressed in the model, so it is not built for governing metrics across hundreds of users. Looker is, and that is the reason companies pay for it.
Does Looker need a data warehouse?
Yes. Looker queries a warehouse and does not store your data itself. Flitch connects directly to 63 sources including databases, files and SaaS applications, so no warehouse is required.
Which is faster to get running?
Flitch, by a wide margin. Looker needs the model written before anyone can build anything, which is typically weeks. Flitch generates a dashboard from a connected source in minutes.