Actyze
Prompt. Act. Analyze.·actyze.ai
Actyze is an open-source, AI-powered analytics platform that converts natural language into SQL queries across multiple databases. It includes a smart semantic layer that learns table relationships, no-code ML pipelines for forecasting and anomaly detection, scheduled KPI materialization, and voice/text queries in 50+ languages. It is self-hosted, running entirely within the user's own infrastructure with federated queries across databases like PostgreSQL, MySQL, Snowflake, and MongoDB.
What it's for
Features 16
- AI-powered schema detection and intent classification
- AI-recommended chart visualizations (bar, line, scatter, pie)
- Air-gapped and on-premise deployment support
- CSV/Excel export
- Federated query engine via Trino without moving data
- Multi-database federation across PostgreSQL, MySQL, Snowflake, MongoDB, BigQuery and 30+ more
- Natural Language to SQL conversion
- No-code ML model training (XGBoost, LightGBM, AutoGluon)
- Predictive Intelligence: forecasting, classification, estimation, anomaly detection
- Query history mining for automatic JOIN discovery
- Scheduled KPI collection and materialization into gold-layer tables
- Schema-only LLM exposure to prevent sending actual data to the model
- Self-hosted deployment via Docker, Kubernetes, Helm
- Smart Semantic Layer with relationship graph and three-layer discovery
- Support for 100+ LLM providers via LiteLLM (Claude, GPT-4o, Gemini, Groq, Together, Ollama, Bedrock)
- Voice and text input in 50+ languages
At a glance
Integrations
Complianceself-reported
Resources
Pricing
The only stated offer is a free, self-hosted open-source deployment (price $0). The site also mentions 'getting a free trial' and 'try for free' language, but no separate paid or hosted pricing tiers are described.
Self-hosted deployment via Docker, Kubernetes, or Helm; AGPL-3.0 licensed with no restricted features.