Local LLMs
Run open models on your own servers. Questions, schemas and results stay on your network.
LLM-Powered Analytics
Your models, inside your walls.
Chat4ED runs on the language model that suits your security rules: a hosted model, a locally deployed open model such as Llama, Qwen or Mistral, or a model fine-tuned on your own data. On-premise and air-gapped deployments never call an outside AI provider.
The problem
Many enterprises are not allowed to send customer, patient or financial data to a third-party model. For them, analytics tools built on a hyperscaler's hosted model are ruled out before the demo.
Fine-tuned for your domain
Tune a model on your terminology, workflows and documents so it understands your business language.
Pay for hardware, not tokens
A local model costs the same however many questions your teams ask, which makes heavy use predictable to budget.
Sample answer
What an answer looks like.
Illustrative example on sample data.
Forecast next quarter revenue and tell me what is driving growth.
Answer 3 of 3 models agreeQ4 is forecast at $15.1M (range $13.9M–$16.3M). Enterprise licences drive 61% of the growth. The model was backtested on the last 4 quarters with 3.2% average error.
SQL it ran
WITH monthly AS (
SELECT DATE_TRUNC('month', order_date) AS month,
SUM(net_revenue) AS revenue
FROM sales.orders
WHERE order_date >= '2024-01-01'
GROUP BY 1
)
FORECAST revenue FROM monthly HORIZON 3;Definitions used
- Revenue = recognised net revenue
- Forecast model: auto-selected, backtested
Sources
sales.orderssales.productsTry asking
Questions teams ask.
“Forecast next quarter revenue and the main drivers”
A forecast with a likely range and its backtest error, computed on your own infrastructure.
“Summarise this month's branch performance in Telugu”
The same verified numbers, explained in the language the user asked in.
“Which customers are most likely to churn?”
A risk score for each customer, with the factors behind every score.
Connectors
Works with the data you already have.
Snowflake
BigQuery
Redshift
PostgreSQL
MySQL
SQL Server
Oracle
ClickHouse
Trino- Athena
SAP HANA & B1
Excel & CSV