Databrain vs Looker Embedded for embedded analytics
Skip the LookML lock-in. Databrain is the developer-friendly embedded analytics platform built for SaaS, with native web components instead of signed iframes and no quote-based contract. Ship in days, scale predictably, pay a flat, published price. See why businesses choose Databrain over Looker Embedded.
Last updated
Short answer
Choose Databrain if
The dashboard is part of your product and has to look like it: Databrain renders as web components in your page, not an iframe, with per-customer themes and white-labeling on every plan.
You run a multi-tenant SaaS and want tenant scope to come from a guest token your backend mints, not from embed users, user attributes and an access filter on every Explore.
You want a published flat price (Growth $999/mo, Pro $1,995/mo) with unlimited seats, embeds and tenants, instead of a quote for an edition plus user licences by role.
Choose Looker if
You want a governed LookML semantic model, kept in Git, as the backbone for both embedded and internal analytics.
You run on Google Cloud and want what Looker (Google Cloud core) adds: IAM, private connections, VPC Service Controls and customer-managed encryption keys. Gemini-based Conversational Analytics is also available on Looker-hosted instances, though not alongside CMEK or VPC Service Controls.
You want customers to explore, not just view: embedded Explores, saved content and downloads, plus a HIPAA Business Associate Agreement and wider certifications.
Keep Looker for internal reporting if it already works there. The two are not mutually exclusive: Databrain can serve the customer-facing layer while Looker stays internal.
A Looker customer moved its customer-facing analytics to Databrain
SpotDraft is the one published Databrain case study that replaced Looker. According to the case study, its Looker data was always one day stale because of an intermediary Google Sheet, and customers had to leave the app for an external link to see reports. That was SpotDraft's setup, not a limit of Looker itself. Figures and the quote come from the published case study, and one customer is a small sample.
A governed BI platform on a LookML semantic model, with an Embed edition for external analytics
How it embeds
Web components (Shadow DOM) for React, Next.js, Angular, Vue, Svelte, Solid and vanilla JS
An iframe that runs the full Looker application, signed per user, with an Embed SDK and JavaScript events
How tenants are scoped
clientId in a server-minted guest token; one data app serves every tenant
Embed users carry user attributes, models and permissions in a signed URL; an access_filter on each Explore; groups for content
Who needs a licence
Unlimited seats on every plan
Platform edition plus Developer, Standard and Viewer user licences by role; no prices published
Pricing model
Flat and published: $999/mo Growth, $1,995/mo Pro, Enterprise custom
Standard, Enterprise and Embed editions each show Call sales; annual terms of one, two or three years
Query model
Live queries against your warehouse
SQL generated from LookML runs against your database
Deployment
Databrain-hosted cloud, or self-hosted on Enterprise
Looker (Google Cloud core) is Google-hosted only; a customer-hosted Looker (original) is a separate instance type
Typical time to production
1 to 4 weeks across our published case studies; SpotDraft 4 weeks
We found no Looker-published timeline
What Databrain and Looker are
Databrain
Databrain is an AI-native embedded analytics platform for B2B SaaS. You connect your warehouse, build metrics and dashboards in the studio, mint a guest token from your backend, and place a web component in your app.
Dashboards are white-labeled, and customers can ask questions in natural language.
Looker
Looker is Google's analytics platform. Analysts describe dimensions, aggregates, calculations and data relationships in LookML, a modeling language usually kept in Git, and business users query that model through Explores and dashboards.
For embedding, Looker offers public, private and signed embedding of dashboards, Looks, Explores and Conversational Analytics in an iframe. It is sold as Looker (Google Cloud core) in Standard, Enterprise and Embed editions, or as Looker (original).
Key differences in one minute
What gets set up for each tenant
In Looker, an embed user is defined by a signed URL carrying an external user ID, permissions, model access, user attributes and optional groups. Row scope comes from an access_filter on each Explore, and Google's docs warn that an Explore without one is unrestricted. For content, Google recommends a closed system with one group per tenant. In Databrain a tenant is the clientId in a guest token, and Databrain's own docs warn that a custom SQL metric must reference the tenant variable or it runs unscoped. Both ask you to get a rule right per object; the objects differ.
How deep the embed goes
Looker's embed puts the whole application in one iframe, which Google describes as more out-of-the-box, and themes let you restyle fonts, colors and hide parts of the page. The default footer reads Powered by Looker, and removing it is a separate feature Google enables. Databrain renders web components in your page, configured through four theming layers. Looker's iframe can expose more of the analyst UI, if you grant the permissions.
Where the modeling lives
Looker's model is LookML, written by analysts, kept in a Git-backed project and reused across every Explore. Databrain metrics are built in a visual builder, with custom SQL or from a question. Our pricing page lists datasets, datamarts and the semantic layer under Enterprise, so confirm what your plan includes.
What the bill tracks
Looker is a quote for an edition plus user licences by role, on annual terms, with monthly API call limits per edition and data-token quotas for Conversational Analytics. Databrain is a flat monthly price by feature, with unlimited seats, embeds and tenants.
Breadth
Looker is the broader platform: Gemini Conversational Analytics on Looker-hosted instances, and private connections, VPC Service Controls, customer-managed encryption keys, Terraform and IAM on Looker (Google Cloud core). Databrain is a narrower embedding layer, which is the point if embedding is what you need.
Feature comparison
Nineteen capabilities that decide a customer-facing analytics build, Databrain against Looker. Looker cells cite Google's own documentation and pricing page, opened on 22 October 2026. Databrain cells cite our pricing page, Trust Center and product docs. Where Looker is stronger, the cell says so.
Databrain vs Looker: 19 capabilities
Capability
Databrain
Looker
Embedding and branding
Yes: Web components in your page
Yes: Signed iframe, with an Embed SDK
Whether the analytics surface is part of your page or an application your page hosts.
<dbn-dashboard> and <dbn-metric> from @databrainhq/plugin render inside your page in Shadow DOM, not in an iframe. An iframe route exists and our docs mark it as not recommended.
Looker's embed puts all the functionality and UI in a single iframe, which Google describes as more out-of-the-box than solutions built from the REST API or visualization components. The iframe runs the whole Looker application and requests only the data your query needs. You can embed publicly, privately (a Looker login is required) or with signed embedding, and use the Embed SDK to create the iframe and message your page.
Yes: Seven framework guides
Yes: Embed SDK for browser apps
How much glue code your team writes per framework.
Guides for React, Next.js, Angular, Vue, Svelte, Solid and vanilla JS, all against the same component. The package declares react and react-dom as peer dependencies, so Angular, Vue, Svelte and vanilla hosts install them too.
The Embed SDK is a library you add to a browser-based web application, available on npm as @looker/embed-sdk, to manage embedded dashboards, Looks, reports and Explores. For signed embedding it needs a backend service that signs the embed URL.
Yes: Four theming layers, per-customer themes
Yes: Embed themes and private label, Enterprise and Embed
How much of your product UI you can restyle, and what removing the vendor footer takes.
Chart appearance, admin theme options, custom chart settings and saved metric settings resolve in a fixed priority order. White-labeling and custom UI themes per customer are on every plan; email white-labeling is on Pro and Enterprise.
Custom themes set the font family, text, background, button and tile colors, and whether headers, titles and the filters bar show, for embedded dashboards and Explores. In Looker (Google Cloud core), themes and private label are in the Enterprise and Embed editions, not Standard. By default embedded content carries a Powered by Looker footer. Removing it is a separate Embed Without Logo feature that Google enables for your instance, with pricing from Google Cloud sales.
Server-minted guest token with a TTL
Signed URL, or cookieless tokens
Customer-facing embeds need an auth path that survives browser cookie changes.
Your backend mints the guest token with your API key and sets its expiry. Every request is checked for a valid token, its expiry, a whitelisted origin and permission.
Your server signs an embed URL with the embed secret, or calls the Create Signed Embed Url API. The URL carries a nonce that cannot repeat within an hour, a timestamp, a session length of up to 30 days, an external user ID, permissions, models and user attributes, and is signed with HMAC-SHA1. Browsers that block third-party cookies need the Cookieless Embed feature, or a Looker host on the same domain as your app.
API tokens, guest-token API, MCP server
Looker API, client SDKs, Terraform
Relevant if you automate tenant onboarding and content changes from code.
Server-to-server API tokens mint guest tokens and manage dashboards and metrics. An Import Dashboard API and an MCP server can provision per-client dashboards from a template.
The Looker API runs on your instance, with client SDKs for non-browser applications and the Embed SDK for the browser. Looker (Google Cloud core) also supports Terraform for some admin tasks. API calls are metered per edition: on Embed, 500,000 query-based and 100,000 admin-related calls a month.
Multi-tenancy and security
Yes: One data app, scoped by clientId
Yes: Embed users, attributes and a closed system
Whether adding a tenant is a value in a token or users, groups and attributes to set up.
One data app serves every tenant. Tenancy can be a shared table with a tenant column, or a schema or database per tenant. Most teams run one workspace, not a workspace per tenant. The pricing page lists Growth as multi-tenant with a single source and Pro as multi-tenant with several.
Each signed embed user has an external user ID, permissions, model access, user attributes and optional groups. For content, Google strongly recommends a closed system (a multitenant installation) for signed embedding, with one group per tenant, so tenants cannot see each other. The Closed System option is requested from Google Cloud sales or support. Google also recommends keeping customer content on a separate Looker instance from internal analytics.
Pass a clientId
Sign a URL with the tenant's attributes
The per-tenant step your onboarding automation has to perform.
A new tenant's data is scoped as soon as your backend mints a guest token with its clientId, provided the tenant column or schema mapping is already set.
Your server signs a URL whose user attributes carry the tenant value that each Explore's access_filter matches. Attributes you set through the API but leave out of the signed URL reset to their defaults the next time the URL is used. In a closed system you also create a group per tenant and set folder access for shared content.
Yes: Tenant filter applied to the query
Yes: access_filter plus user attributes
Both are real row-level security. The difference is where the rule lives and what you must remember to apply on each object.
For builder metrics Databrain injects the tenant filter into the generated query server-side. Per-session constraints go in rlsSettings in the guest token. Custom SQL metrics are not auto-scoped: the author must reference the tenant variable, or the metric runs unscoped. Row and column level security rules are listed under Enterprise on the pricing page.
An access_filter ties an Explore field to a user attribute and must be added to every Explore that needs it: Google's docs warn that an Explore without one leaves data unrestricted. Every user needs a value for the attribute, even admins. For embed users the values come from the user attributes in the signed URL.
Partial: SOC 2 and GDPR; ISO 27001 in progress
Yes: Certifications, HIPAA BAA availableLooker leads here
If a buyer's procurement list names HIPAA with a BAA or FedRAMP, Looker documents a route and Databrain does not yet.
Databrain's Trust Center lists SOC 2 and GDPR as compliant and ISO 27001 as in progress, and our security documentation describes the SOC 2 report as Type 2. HIPAA is listed as coming soon on Pro and Enterprise.
Google says Looker products maintain several ISO, SOC and other certifications, listed on its services-in-scope page, and that a HIPAA Business Associate Agreement is available. For Looker (Google Cloud core), compliance certifications such as FedRAMP High are available only on the Enterprise and Embed editions.
Yes: Cloud, or self-hosted on Enterprise
Yes: Google-hosted, private networking on higher editions
Regulated customers ask where the data sits and who can reach it.
Databrain-hosted cloud, or self-hosted in your own infrastructure on Enterprise, with VPC peering and a bastion host. Queries run live against your database, and our docs state that Databrain stores only metadata such as dashboard configurations and metric definitions, while your data stays in your database.
Looker (Google Cloud core) is hosted by Google in Google Cloud and is not available for customer-hosted or multicloud environments. Private connections, VPC Service Controls and customer-managed encryption keys are in the Enterprise and Embed editions, and Google lists US, UK, Hong Kong and Korea as hosting locations for it. Looker (original) is a separate instance type that can be Looker-hosted or customer-hosted.
Data and semantic layer
Partial: Metrics and calculated fields; semantic layer on Enterprise
Yes: LookML semantic model, kept in GitLooker leads here
Governed metric definitions reused across every dashboard are Looker's core strength.
Databrain builds metrics on your connected data and adds calculated fields. On our pricing page, datasets, datamarts and the semantic layer sit in the Data studio group and are listed under Enterprise, so confirm what your plan includes.
LookML describes dimensions, aggregates, calculations and data relationships once, and Looker generates SQL from it, so business users can query without writing LookML or SQL. A LookML project is a set of files typically version-controlled through a Git repository.
Builder, custom SQL, calculated fields
LookML
Who on your team can ship a new metric without help.
Metrics are built in a visual builder, with custom SQL, or from a natural-language question. Calculated fields add derived values without editing the source data.
LookML is a dependency language, like make, and is independent of particular SQL dialects. Google notes that writing LookML requires an understanding of SQL queries, though you do not have to be a SQL expert.
Yes: Warehouses and databases, live
Yes: Long dialect list
Confirm your exact warehouse in both lists before you commit.
Snowflake, BigQuery, Redshift, Databricks, ClickHouse, Postgres, MySQL, MSSQL, Athena, Trino, MongoDB, Elasticsearch and more. Growth includes one data source; Pro and Enterprise are unlimited.
Looker lists Snowflake, Google BigQuery, Amazon Redshift, Databricks, ClickHouse, PostgreSQL, MySQL, SQL Server, Athena, Trino, Oracle, Teradata, SAP HANA and others as supported dialects. Google notes that because LookML builds on in-database features, some dialects allow a more powerful implementation than others. The list we read is for Looker (original); Looker (Google Cloud core) publishes its own.
Yes: Live queries; caching on Pro and up
Yes: Live SQL generated from LookML
Both query your warehouse rather than copying it, so freshness depends on your own pipelines and caching.
Every surface queries your warehouse live through the same path. Caching is listed for Pro and Enterprise. CSV uploads are snapshots, not live connections.
Looker's SQL generator translates LookML into SQL queries that run against your database. Google says there is no permanent storage of customer data in your Looker instance, and that cached query data is encrypted with AES-256.
Partial: BigQuery connector
Yes: IAM, private connections, VPC-SC, CMEKLooker leads here
If your stack and security model are Google Cloud, Looker (Google Cloud core) plugs into it directly.
Databrain connects to BigQuery like any other warehouse. Our docs do not describe integration with Google Cloud IAM, VPC Service Controls or customer-managed encryption keys.
Looker (Google Cloud core) integrates with Google Cloud IAM and BigQuery connectivity. On Enterprise and Embed it adds private connections, VPC Service Controls and customer-managed encryption keys. It also offers Cloud Audit Logs and, for BigQuery sources, data lineage through Knowledge Catalog.
What your customer gets
Partial: End-user report creation on Enterprise
Yes: Embedded Explores, saving and downloadsLooker leads here
If customers should explore freely instead of opening fixed dashboards, Looker's embed can expose full Explore pages.
Natural-language search is on every plan and follow-up chat on Pro and Enterprise. End-user report creation is an Enterprise feature, and the workspace and permissions must also allow it.
Signed embedding can expose Explores, and permissions let an embed user create custom fields and table calculations, save content, see the SQL, download results, schedule deliveries and create alerts. Embedded themes can hide parts of the Explore page.
Yes: Natural-language search on every plan
Yes: Conversational Analytics, built on Gemini
Both answer questions in natural language. They differ in which model you use and in how the bill treats usage.
Natural-language search is included on every plan, and follow-up chat on Pro and Enterprise. Connect OpenAI, Azure OpenAI, Claude, Gemini, Llama or Mixtral. Questions run through the same tenant-scoped query path as dashboards.
Conversational Analytics can be embedded with private or signed embedding, needs a Looker-hosted instance, and embed users need permissions such as chat_with_explore. Embedded themes only partly apply to it. Usage is measured in data tokens against a monthly quota per edition. Google says access is currently promotional, with no quota limits or overage fees within fair usage, and will give at least 90 days' notice before enforcement, after which overage is $3.00 per 1M input tokens and $20.00 per 1M output tokens. Gemini in Looker is not supported on Looker (Google Cloud core) instances that use customer-managed encryption keys or VPC Service Controls.
Commercial
Yes: Flat and published
Call sales; edition plus users by role
One of these bills is a number you can read today; the other starts with a quote.
Growth $999/mo, Pro $1,995/mo, Enterprise custom. Seats, embeds and tenants are unlimited on every plan.
The Standard, Enterprise and Embed editions each show Call sales. Each platform includes one production instance, 10 Standard users and 2 Developer users, and further Developer, Standard and Viewer licences are added by role, on annual terms of one, two or three years. API calls per month are capped by edition. Google publishes no amounts.
14-day trial; POC on your data for Pro and Enterprise
90-day trial instanceLooker leads here
A longer trial gives more time to test a real embed.
14-day free trial on every plan, with a proof of concept on your own dataset for Pro and Enterprise. Enterprise adds 24x7 premium support, a 99.5% uptime SLA and a dedicated account manager.
Google lets you create a trial Standard, Enterprise or Embed instance with the same feature support as the production edition. Trial instances last 90 days, and upgrading means creating a new instance and importing from the trial. Non-production instances for staging are purchased separately.
How multi-tenancy and security work
Both let you embed for many customers without giving each a login to the BI tool. They differ in where tenant scope lives: Looker puts it in the embed user's attributes and the LookML model, and Databrain puts it in the guest token and the generated query.
Databrain
Your backend mints a guest token that carries the tenant clientId. The embed sends that token with each request. Databrain checks the token, its expiry, the whitelisted origin and permission, then applies the tenant filter to the query before it reaches your warehouse. One data app serves every tenant, and extra per-session constraints go in rlsSettings.
One sharp edge: in custom SQL metrics the tenant filter is not auto-injected, so the author must reference the tenant variable. A custom metric that omits it runs unscoped.
Databrain's Trust Center lists SOC 2 and GDPR as compliant and ISO 27001 as in progress, with HIPAA coming soon. Compare those lists against your buyers' requirements.
Looker
Signed embedding authenticates users through your own application. Your server signs a URL that carries an external user ID, permissions, model access, user attributes and optional groups, and the user attributes drive the access_filter on each Explore.
Row-level security is applied per Explore. Google's docs warn that an Explore without an access_filter leaves data unrestricted, and that every user, even an admin, needs a value for the attribute.
For content, Google strongly recommends a closed system with one group per tenant, requested from Google Cloud sales or support, and recommends keeping customer content on a separate Looker instance from internal analytics.
Two cautions from Google's own docs: an external user ID must be unique for each set of permissions, attributes and models, or data can become visible to users who should not see it, and the embed secret should be protected like an admin credential, because anyone holding it can sign a URL for any model, as any user.
A Databrain request
Your backend calls the guest-token endpoint with the customer's clientId and your data app name.
POST /api/v2/guest-token/create
Your app passes the token to the web component.
Databrain checks the token, its expiry, the origin allow-list and permission.
The tenant filter is applied to the query.
The query runs against your warehouse and the result is returned to the component.
A Looker request
Your server signs an embed URL, or calls Create Signed Embed Url, with an external user ID, permissions, models, user attributes and a session length.
The browser loads the signed URL in an iframe, directly or through the Embed SDK.
Looker verifies the signature and starts a session for that external user. A signed URL that is not used within five minutes expires.
Explores with an access_filter take the filter value from the user's attribute.
The query runs against your database and the result renders in the iframe.
Both are real tenant isolation; the difference is whether a tenant is an embed user with attributes in Looker's model or a value in a token.
When each is a better fit
Both lists are genuine. If your situation matches the Looker side, that is the right call.
Databrain is the better fit when
The analytics is a feature of your product and has to match its design.
You are multi-tenant and onboarding tenant 500 should be a token, not another embed user definition to sign.
Finance needs a number it can forecast from a published price.
You want metrics built in a visual builder, with custom SQL or from a question, without a modeling language.
You want natural-language questions on every plan with your choice of LLM.
You need self-hosting, VPC peering or a bastion host (Enterprise).
Looker is the better fit when
You want one governed LookML model behind embedded and internal analytics, and your team is ready to model in it.
You run on Google Cloud and want IAM, private connections, VPC Service Controls and customer-managed encryption keys.
You want customers to explore data in embedded Explores, save content and download results.
You need a HIPAA Business Associate Agreement or certifications beyond the ones Databrain lists.
You want Gemini-based Conversational Analytics.
You can budget a quoted annual contract and want a 90-day trial instance.
Decision guide
Match your situation to the row, then check the feature comparison for the detail.
Which product fits which situation
If your question is
Lean toward
Because
The viewer is your customer, not your colleague
Databrain
Customer-facing embedding, per-customer themes and no per-viewer cost are what Databrain is built for.
Many tenants with self-serve onboarding
Databrain
Tenant scope comes from a guest token, and one data app serves all tenants. Looker can automate user creation through its API, but each tenant still needs attributes, and in a closed system a group.
The dashboard must match your product UI
Databrain
Web components and four theming layers. Looker has embedded themes and private label too, so check both against your design system.
You need a governed semantic layer shared across tools
Looker
LookML in Git is Looker's core, and Databrain's modeling is simpler.
Your stack and security model are Google Cloud
Looker
Looker (Google Cloud core) adds IAM, private connections, VPC Service Controls and customer-managed keys on Enterprise and Embed.
Customers must explore and save their own content
Looker
Signed embedding can expose Explores with save and download permissions. Databrain's end-user report creation is Enterprise.
You need HIPAA with a BAA, or FedRAMP
Looker
Google says a BAA is available, and lists FedRAMP High for Looker (Google Cloud core) Enterprise and Embed. Databrain lists HIPAA as coming soon.
You need forecastable spend from a published price
Databrain
A flat monthly plan with unlimited seats, embeds and tenants.
You need to self-host the analytics layer
Either
Databrain offers it on Enterprise. Looker documents a customer-hosted Looker (original); Looker (Google Cloud core) is Google-hosted only.
Pricing snapshot
A flat plan and a quote for an edition plus users by role are different shapes of bill, and Google publishes no Looker amounts, so compare Databrain against a quote at your own tenant and user counts. Neither is cheaper in every case.
Databrain
Growth
$999/mo
Pro
$1,995/mo
Enterprise
Custom
Seats, embeds, tenants
Unlimited
Free trial
14 days
Growth includes multi-tenancy and 1 data source. Pro adds unlimited data sources, SSO and roles and permissions. Enterprise adds self-hosting, multi-region deployment, VPC peering and bastion host, and a dedicated account manager.
Looker
Standard edition
Call sales
Enterprise edition
Call sales
Embed edition
Call sales
Contract
Annual, 1 to 3 years
Trial
90-day instance
Each edition includes one production instance, 10 Standard users and 2 Developer users, and further Developer, Standard and Viewer licences are added by role, with no prices published. Query-based API calls per month are capped at 1,000 on Standard, 100,000 on Enterprise and 500,000 on Embed. In Looker (Google Cloud core), signed embedding and cookieless embed are only in the Embed edition. Google recommends customer content live on a separate instance from internal analytics, and a non-production instance for staging is purchased separately. Conversational Analytics is measured in data tokens with a monthly quota per edition. Access is currently promotional, and Google lists overage of $3.00 per 1M input tokens and $20.00 per 1M output tokens once enforcement starts, with at least 90 days' notice.
Google's pricing page does not state how external embed users are counted against user licences, so ask for that in your quote. Databrain prices are from the pricing page. Both were read on 22 October 2026.
Keep Looker for internal reporting if it works there. The move is for the customer-facing layer. Databrain's published case studies went live in 1 to 4 weeks, and you can run both side by side until you cut over.
Inventory what customers open
List the dashboards and Explores customers use most, and the LookML dimensions and measures behind them.
Connect your warehouse
Connect to Databrain with read-only access.
Rebuild the metrics
Recreate the measures and dimensions you need as Databrain metrics, using the visual builder, custom SQL or natural language.
Recreate the priority dashboards
Rebuild them in the studio and match them to your product design.
Set the tenant mapping and mint guest tokens
Replace the signed URL and its user attributes with a token your backend mints with each customer's clientId.
Pilot with a few tenants
Replace the Looker iframe with a Databrain web component for a pilot group and run both.
Cut over
Move the remaining tenants, then retire the Looker embed users and groups you no longer need, and keep signed embedding disabled if you stop using it, as Google advises.
Plan to recreate dashboards rather than import them: we found no documentation of a Looker or LookML importer. SpotDraft is the one published Databrain customer that replaced Looker, and it took 4 weeks. Databrain's five published case studies went live in 1 to 4 weeks, and SpotDraft's 4 weeks is the longest of them.
Yes, for customer-facing analytics inside a SaaS product. Databrain renders as web components, scopes each tenant from a guest token and charges a flat monthly plan, and SpotDraft moved from Looker to Databrain in 4 weeks. It is not a replacement for Looker as an internal governed BI platform on a LookML model, so keep Looker for that if you rely on it.
How does Looker pricing compare with Databrain pricing?
Google publishes no Looker amounts: the Standard, Enterprise and Embed editions each show Call sales, with user licences by role added and annual terms of one to three years. Signed embedding is only in the Embed edition of Looker (Google Cloud core). Databrain is flat and published: $999 per month for Growth and $1,995 for Pro, with unlimited seats, embeds and tenants. Ask Google for a quote at your tenant and user counts.
Does Looker Embedded support multi-tenancy?
Yes. Google recommends a closed system, which it also calls a multitenant installation, with one group per tenant. Your server signs an embed URL carrying the tenant's user attributes, and an access_filter on each Explore uses them to restrict rows. In Databrain you pass a clientId in a guest token and one data app serves every tenant. Looker gives more control over content access; Databrain has fewer objects to set up per tenant.
Do I need to learn LookML to use Databrain?
No. Databrain metrics are built in a visual builder, with custom SQL or from a natural-language question. LookML is Looker's modeling language, and Google notes that writing it requires an understanding of SQL. If your team already models in LookML, that investment is a point in Looker's favor, since the semantic model is what Looker is built around.
Can I white-label both?
Yes. Databrain includes white-labeling and per-customer themes on every plan. Looker offers embedded themes and a private label option on its Enterprise and Embed editions, and removing the Powered by Looker footer from embeds is a separate feature that Google enables for your instance. Compare both against your design system before you decide.
Can I self-host Looker or Databrain?
Databrain offers self-hosting on Enterprise, with VPC peering and a bastion host. Looker (Google Cloud core) is hosted by Google and is not available for customer-hosted or multicloud environments. Looker's documentation describes a separate customer-hosted Looker (original) instance type, so ask Google which one applies to your contract.
Does Looker have AI for end users?
Yes. Looker's Conversational Analytics is built on Gemini and can be embedded with private or signed embedding, with usage measured in data tokens against a monthly quota. Databrain includes natural-language search on every plan and follow-up chat on Pro and Enterprise, with your choice of LLM connector. Compare how each prices usage before you decide.
How long does switching from Looker take?
Databrain's published case studies went live in 1 to 4 weeks. SpotDraft replaced Looker in 4 weeks, according to its published case study. Treat that as one data point: we found no Looker-published migration timeline, and every LookML model and set of embed users is different. Plan to rebuild dashboards, because we found no documentation of a Looker importer.
Methodology and sources
We check every Looker claim against Google's own public documentation and pricing page, and we date the check: every Looker source below was opened on 22 October 2026. Google now serves Looker documentation from docs.cloud.google.com, and cloud.google.com/looker/docs addresses redirect there.
Looker Embedded is how buyers search, but Google sells Looker (Google Cloud core) in Standard, Enterprise and Embed editions, and Looker (original) separately. Where a feature depends on the edition or instance type, the cell says so, and most claims come from the shared Looker docs or the Looker (Google Cloud core) docs.
Databrain claims come from our published pricing page, Trust Center, product documentation and customer case study. Google publishes no Looker prices, so we do not state a Looker price or implementation time. Where we found no Looker documentation of a capability, the cell says so, not that it is impossible. We did not compare Looker support terms or the named certifications behind Google's compliance pages.
Where Looker is stronger, we say so. If you find something here that is out of date or wrong, tell us and we will fix it.
Looker (Google Cloud core) overviewGoogle Cloud docs, opened 22 October 2026. Source for hosting, trials, non-production instances and the Looker (original) comparison
Signed embeddingGoogle Cloud docs, opened 22 October 2026. Source for the URL parameters, permissions, cookieless embed and the closed system recommendation
Thirty minutes. Bring your architecture and your warehouse, and we will show a white-labeled dashboard running against your schema and tell you plainly if you are a bad fit.
