What size of company is this solution suitable for?

From startups to large enterprises. The free plan suits individuals and small teams, Pro is for professional analysts, Team and Enterprise plans are for medium to large organizations.

Can it integrate with our existing database systems?

22+ data sources are supported, including PostgreSQL, MySQL, BigQuery, Snowflake and more. Connect directly to your existing databases — no data migration required.

Can non-technical users really use it?

Absolutely. Our core focus is enabling business users without SQL skills. Simply describe your analysis needs in plain language, and the AI automatically generates SQL and returns visualized results.

Data Analytics Solution

Break down data analytics barriers with AI — empower every business user to drive data-driven decisions

Why AI-Powered Data Analytics?

Data Analytics Democratization

Eliminate the SQL skill barrier, enabling every business user to perform data analysis independently

Faster Decision Making

From request to insight — reduced from days to seconds

Data Team Empowerment

Reduce repetitive data requests by 80%, freeing data teams for high-value analysis

Unified Data Semantics

MDL semantic modeling ensures consistent business definitions and calculation logic across the organization

Use Cases

> Retail: Sales trend analysis, product affinity analysis, inventory optimization

> Finance: Risk management reports, customer profiling, compliance auditing

> Manufacturing: Production quality analysis, supply chain optimization, predictive maintenance

> E-commerce: User behavior analysis, conversion funnel, campaign effectiveness

> SaaS: Customer health analysis, churn prediction, product usage analytics

> Healthcare: Patient data analysis, clinical research statistics, operational efficiency

FAQ

Make data analytics 'something everyone does daily'

The last decade of BI got us as far as 'the data team makes nice-looking reports'. The remaining problem: when a business user wants to drill in or slice a dimension, they're back in Excel or bothering the data team. Our data analytics solution closes that gap. MDL maps physical tables to business terms; the LLM turns one Chinese/English sentence into safety-checked SQL; results stream back with charts. Business users don't wait, don't beg, don't learn SQL. The data team goes back to building assets instead of running a ticket funnel.

Three real scenarios

Retail chain

Situation · HQ wants East-China trend + top-10 SKUs + WoW; five departments ask five slightly different versions daily.

Outcome · 'Revenue/margin/AOV' definitions live in MDL Cubes; a business user asks, AI returns the same-definition SQL with a WoW-tagged chart in 30 seconds.

Financial risk

Situation · Compliance audits pull anomalous transactions across databases with row/column permissions; DBAs won't issue ad-hoc accounts.

Outcome · Strict SQL mode rejects writes; row/column permissions bind to RBAC; auditors see only authorized rows; all queries/SQL are logged and replayable.

Manufacturing

Situation · Production data lives in MES, SCADA, ERP; quality engineers stitch it together daily with slippery definitions.

Outcome · Yitu connects the three, MDL defines yield/downtime/work-order as business terms, engineers ask in natural language and slice by shift/line for immediate on-floor review.

Performance and scale

22+Data sources

PG/MySQL/BigQuery/Snowflake/ClickHouse/Databricks/Trino etc.

95%+SQL accuracy

Blind test with MDL + semantic memory enabled

80%↓Ticket drop

Data-team ticket volume after 3 months on Yitu

<3 secMedian query response

From prompt to first SSE chunk

1 / metricDefinition uniqueness

One MDL definition per metric company-wide

≤2 weeksTime to value

Connect + MDL + train business users to self-serve

How we compare

vs traditional BI (Tableau / PowerBI / FanRuan)

Natural-language queries; MDL keeps definitions unified.

Them

Drag fields, write DAX; scattered logic; business users still depend on data.

Takeaway · Complementary: BI for deep reports, Yitu for daily self-serve.

vs generic ChatBI / Text2SQL OSS

MDL + semantic memory + safety policy; enterprise-grade access native.

Them

Usually schema-in-prompt; accuracy collapses on complex JOINs and permission constraints.

Takeaway · Demo-to-production gap is exactly the semantic layer and safety policy.

vs building on LLM + LangChain

SaaS or on-prem, value in 2 weeks, upgrades by platform.

Them

Selection to alerting, all self-built; hidden long-term maintenance.

Takeaway · First-time working is easy; three years without incidents is the real cost.

Deeper details

Why MDL matters

MDL is 'business-definition source code' — defining 'revenue', 'order', 'customer' to the field level. The LLM reads MDL before generating SQL, eliminating the classic 'same metric, different answers' problem. Data teams keep their expertise; business users get autonomy — both sides do what they're better at.

SQL safety policies

All DDL/DML writes blocked by default; SELECTs pass RBAC checks with row/column controls injected before dispatch; credentials Fernet-encrypted and decrypted only at execution. Every query and SQL logged, replayable per user/role/time — supports GB/T 22239 level 3 and financial-sector audits.

Industry templates and best practice

For retail, finance, SaaS, manufacturing, e-commerce, and education, Yitu ships MDL templates and common Cube definitions; customers fine-tune and go live in 2 weeks. These templates are one of our strongest differentiators against generic tooling.

Rollout path

Week 1: connect DB, load MDL skeleton, run 3-5 high-frequency questions. Week 2: train business users, roll out in one department. Month 1: SSO/RBAC, expand to 3+ departments. Month 3+: retire 'data request' tickets, data team returns to asset building.

Start Your Data-Driven Journey

Empower every business user to perform data analysis independently

Get Started Free 联系我们