feldera/feldera

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

2064 stars · 149 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Feldera is an incremental computation engine written in Rust that evaluates arbitrary SQL programs incrementally using DBSP theory, maintaining materialized views over changing data. It processes inserts, updates, and deletes to SQL tables and updates all dependent views without recomputing from scratch, supporting full SQL semantics including joins, window functions, and recursion.

Use cases

• keep SQL views fresh over streaming data with incremental view maintenance
• build real-time ETL pipelines in SQL
• run feature engineering pipelines on live and historical data
• power real-time dashboards with millisecond updates
• process datasets larger than RAM incrementally
• replace batch warehouse jobs with continuous incremental analytics

When to choose

• you need full SQL semantics (joins, aggregates, window functions, recursion) evaluated incrementally
• you want unified batch and streaming compute in one pipeline
• you need high-throughput incremental view maintenance without manual tuning
• your datasets exceed available memory

When to avoid

• you need a full transactional OLTP database - Feldera is a compute engine, not a database
• you need multi-node fault-tolerant deployments on the open-source edition (that requires the paid Enterprise tier)
• you only run simple periodic batch jobs where incremental computation adds no benefit

Facets

service · maturity active

database streaming etl analytics search-engine databases analytics big-data rust self-hosted cloud incremental-view-maintenance sql stream-processing dbsp materialized-views data-pipelines real-time-analytics data-engineering real-time docker linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="feldera/feldera")