Qrly vs Sigma
Sigma is spreadsheet-shaped cloud BI: SaaS only, with a cloud data warehouse required underneath it and no on-prem option for the platform or its AI. Qrly is BI + AI in one self-hostable platform — natural-language Ask (NL→SQL), anomaly detection and schema descriptions all running on your choice of local LLM (Ollama, LM Studio) or cloud (Anthropic Claude, Google Gemini, OpenAI, Azure OpenAI), across 40 connection types, with flat pricing for unlimited users.
Who wins where
No marketing fluff. Sigma is a well-regarded product with a long track record — here is where each tool is genuinely stronger, and where teams consistently land during evaluation calls.
BI + AI in one self-hosted platform, flat pricing
- AI Ask (NL→SQL) included — on your own hardware if you want
- BYO LLM: local Ollama / LM Studio, or Claude / Gemini / OpenAI / Azure
- Embedded analytics with signed JWT — no per-viewer seats
- AI anomaly detection, schema descriptions, performance analyzer built-in
- Flat licence for unlimited users — no seat-type arithmetic
- Multi-tenant hierarchy, OIDC and LDAP sign-in as standard
Self-service analysis for people who do not write SQL
- Both let non-analysts build their own questions without SQL
- Both query the database live rather than a stale extract
- Both schedule results out to people on a cadence
- Both expose a REST API and embed into your own application
The spreadsheet grid, and the warehouse under it
- A spreadsheet interface non-analysts pick up in minutes
- Input tables and writeback into the warehouse
- Live query at warehouse scale, with no extract to manage
- A polished, fully managed cloud service — nothing to run
Feature-by-feature comparison
The features most teams actually evaluate when comparing Qrly and Sigma.
| Feature | RecommendedQrlySelf-hosted · Belgium | SigmaSigma Computing |
|---|---|---|
| Self-hostable on your own infra | Included | SaaS only, no self-hosting |
| No per-user licence counting | Included | Priced per user by account type |
| Built-in customer embed portal | Included | Sigma Embedded, licensed separately |
| Alerts on saved questions (email, Slack, webhook) | Included | Scheduled alerts on workbooks |
| Native scheduled subscription (4 providers) | Included | Scheduled exports |
| Visual query builder and raw SQL, round-trip convertible | SQL ⇄ visual | Spreadsheet generates SQL one-way |
| Azure AD + Google + LDAP + Basic simultaneously | Included | SAML and OAuth, no LDAP |
| OIDC SSO user provisioning | Included | SAML SSO, plan-dependent |
| AI with on-prem option (Ollama, LM Studio) | Included | Cloud AI only, no local LLM |
| Multi-tenant architecture out of the box | Included | Row-level security and workspaces, no tenant hierarchy |
| 40 connection types, 12 SQL dialects | Included | Warehouse-first connector list |
| Flat pricing (unlimited users) | Included | Per user, plus warehouse compute |
| Productive in under 5 minutes | Included | Quick once a warehouse is connected |
| EU data residency (native, not a tier) | Included | Vendor cloud regions only |
| No add-on purchase required for the basics | Included | Embedding is licensed separately |
| Dashboards with cascade filters and live push | Included | Dashboards and filters, no push |
| REST API + webhooks | Included | Included |
| OLAP models with ROLLUP, CUBE, prior-period | Included | No CUBE or prior-period columns |
| Runs on your own database, no warehouse needed | Included | Cloud data warehouse required |
Four reasons teams leave Sigma
From real migration conversations with data teams and the people who own the platform bill.
Embedding is a separate line on the invoice
Sigma embeds well, and embedding is a separate commercial motion with its own terms sitting alongside the internal licence rather than inside it. What starts as a request from your product team turns into a procurement exercise, and the shape of that contract quietly starts deciding how many customers you are willing to put a dashboard in front of.
Qrly's embedding is the same surface as everything else. Sign a JWT, drop in an iframe, pass locked parameters the viewer cannot override, and serve as many customer tenants as you like — no per-viewer seat and no second contract. The dashboard your customer sees is the one your analyst built, governed by the same permissions and the same always-applied model filters.
Two bills, and only one of them is Sigma's
Sigma is quoted per user by account type, which on its own is predictable enough. The part that catches finance out sits underneath it: every workbook runs live against Snowflake, BigQuery or Redshift, so each refresh is warehouse compute billed by your cloud provider. Give a dashboard to two hundred people who reload it each morning and you have created a recurring second bill nobody signed off.
Qrly's flat licence never scales with headcount — the 51st user costs nothing, and neither does the 501st — and four opt-in caching tiers, result materialisation on a refresh interval, a per-user daily query budget and a query governor exist precisely so a popular dashboard does not turn into a variable cost. Budget conversations become a one-line item.
AI you cannot point at your own hardware
Sigma's AI runs where Sigma runs: in the vendor's cloud, against models the vendor chose. The quality is good, but for regulated teams, legal, finance or public-sector operators the data-flow shape is often the whole decision, because the prompts and the schema they carry still leave the perimeter — and on a SaaS-only platform there is no configuration that changes it.
Qrly ships AI with a local option: point it at an Ollama or LM Studio endpoint on your own hardware and every prompt, every schema description and every generated query stays inside your network. Row access is off by default — the model sees your schema, not your rows, until you explicitly enable it — and there is no separate AI subscription to renegotiate when a sensitive project starts.
SaaS-only is a decision you make once
Sigma is a good product with one structural constraint: it is a cloud service over a cloud warehouse, and both halves are somebody else's infrastructure. If the data cannot leave the building, if procurement needs the platform inside your own network, or if your source of truth is a Postgres or SQL Server instance rather than a warehouse, the evaluation ends there no matter how good the interface is.
Qrly self-hosts as a single deployment on any Linux box or Kubernetes cluster, speaks 12 SQL dialects across 40 connection types, and holds nothing back for a cloud tier. Your data stays where it already is, your LLM stays where you put it, and there is no phone-home in the middle.
Pricing, honestly
Our own list price in full, and an honest account of how Sigma charges. The Qrly column is the published schedule evaluated at €15M of revenue — the number of people using it does not move it.
Methodology, 28 August 2026. Qrly publishes a complete rate card, so the figures in our column are quoted exactly. The other column describes a pricing model rather than quoting a number: vendor list prices change without notice and several of these vendors do not publish one at all. For current figures, go to the vendor’s own pricing page.
Qrly
Annual licence · priced on your revenue · unlimited users, one organisation, one installation
- Annual licence at €15M revenue €6,300
- Per analyst / per viewer €0
- Per query, per credit, per capacity unit €0
- Embedded dashboard consumers €0
- Embedding, alerts, OIDC/SSO, row-level rules Included at every tier
- Self-hosting and bring-your-own LLM Included at every tier
Three years of licence: €18,900
That is our published schedule evaluated at €15M of revenue — 0.18% of the first €500k, then 0.12%, 0.0525% and 0.027% on each tranche above it. It does not move when you add users. The whole rate card, from the €900 floor upward, is on the pricing page.
Sigma
Per user, by role · quoted · Sigma cloud only
- How it charges Per user, by role, on an annual agreement
- List prices Not published — quote only, see sigmacomputing.com/pricing
- Deployment Sigma cloud only — no self-hosted edition
- Embedded analytics Its own commercial arrangement
- Warehouse compute Billed by Snowflake, BigQuery or Redshift
- AI features Vendor-hosted models
Two bills, and Sigma quotes only one of them. The licence is per user and negotiated; the compute is charged by your warehouse on every refresh, every filter change and every embedded view, and that meter answers to your users’ curiosity rather than to your turnover. Model the second bill before you argue about the first.
Moving from Sigma to Qrly
Most teams are up and running on Qrly within a working week.
The standard migration path
There is no export to negotiate. Qrly connects to the same warehouse Sigma queries — Snowflake, BigQuery and Redshift are all supported connection types — so the data never moves and the work is a rebuild of definitions, not a transfer of rows. Mid-size teams typically finish the cutover in under a week; larger organisations run a parallel window of thirty days before dropping the old contract.
- Connect the warehouse directly. Add Snowflake, BigQuery or Redshift as a Qrly connection, test it before saving, then run a schema sync so the visual builder and the SQL autocomplete know your tables. Credentials are AES-encrypted at rest, execution limits are set per connection — timeout, max rows, concurrency, queue depth — and the pool hands out read-only connections.
- Rebuild workbook logic as questions. Each element becomes a saved question: a source table or another question compiled as a subquery, a nested AND/OR filter tree over around 35 operators, aggregations, breakouts, having and order-by with explicit NULL handling. Spreadsheet formulas doing arithmetic on columns become calculated fields, validated server-side before they ever run.
- Promote the repeated logic into a model. Where several workbooks recompute the same measures, define one OLAP model — fact table, dimension joins, a measure and dimension catalogue, named hierarchies — and every question above it inherits the definitions, the year-to-hour time drill, prior-period comparison and the always-applied model filters that act as a security boundary.
- Rebuild dashboards and re-point the embeds. Cards drop onto a 12-column grid with cascade filters mapped per card, and the customer-facing views come back as signed-JWT embeds with locked parameters — the same surface, without a per-viewer seat behind it.
- Move users, then run in parallel. Point Qrly at your identity provider — OIDC and LDAP are in the base product, so nobody re-registers. Keep Sigma live for thirty days while the team compares numbers side by side, and stop renewing once the last verified question has been signed off.
Common questions
Is Qrly a direct replacement for Sigma?
How do we move our Sigma workbooks across?
Our people live in spreadsheets. Will they cope with Qrly?
Does Qrly have anything like Sigma input tables?
How does QQL compare to Sigma's spreadsheet interface?
Cost for 50 users over 3 years?
The full alternatives matrix
Evaluating Qrly against another tool? Jump to the detailed side-by-side.