Real-time analytics is a class of analytical workload where data is queried within seconds of being produced and results are returned with low enough latency to be consumed by an API or a user-facing interface. Read full definition

OLAP (Online Analytical Processing) is a database workload shape characterized by large, read-heavy aggregation queries over historical or event data, as opposed to OLTP, which handles small transactional reads and writes. Read full definition

A column-oriented database stores each column of a table contiguously on disk, instead of storing each row contiguously, which makes analytical aggregations dramatically faster and more compressible than in row-oriented databases. Read full definition

ClickHouse is an open-source, column-oriented OLAP database designed for high-throughput analytical queries over very large datasets, and is the engine Tinybird is built on. Read full definition

A time series is a sequence of data points indexed by time, where the timestamp is the primary axis of analysis and queries are typically expressed as aggregations over time windows. Read full definition

User-facing analytics is analytical data served directly inside a product UI, where end users (not internal analysts) trigger the queries and expect API-grade latency, multi-tenant isolation, and high concurrency. Read full definition

Streaming ingestion is the practice of writing events into an analytical database as they happen, one small batch at a time, so that queries can return results within seconds of the underlying activity instead of waiting on a nightly ETL window. Read full definition

A materialized view is a precomputed query result stored on disk, refreshed as new data lands, so that read queries return aggregated values in milliseconds instead of scanning the raw table. Read full definition

Schema evolution is the practice of changing a database schema (columns, types, indexes, materialized views) without downtime and without breaking the producers or consumers that depend on it. Read full definition

A data source in Tinybird is the typed, append-only table where ingested events land, with a defined schema, sort key, and ingestion connectors attached to it. Read full definition

A pipe is the unit of SQL composition in Tinybird: a sequence of named nodes, each one a query, that can be published as an endpoint, materialized into a data source, scheduled as a copy, or pushed to an external sink. Read full definition

An endpoint in Tinybird is a published SQL pipe exposed as an HTTPS API, with query parameters, authentication, and a JSON response, ready to be called from any application. Read full definition

A copy pipe is a Tinybird pipe that runs on a schedule and writes its result into another data source inside the same workspace. Read full definition

A sink pipe is a Tinybird pipe that pushes query results out to an external destination on a schedule, instead of returning them through an API. Read full definition

Query parameters are typed, validated placeholders inside a Tinybird pipe that turn a static SQL query into a parameterized HTTP endpoint, where each parameter becomes a query string argument with a declared type and optional default. Read full definition

A workspace is the top-level container in Tinybird: a self-contained environment with its own data sources, pipes, endpoints, tokens, members, and resource limits, scoped to a single project or environment. Read full definition

Connectors are managed integrations that move data from external systems (Kafka, S3, BigQuery, Postgres, and others) into Tinybird data sources, on a continuous or scheduled basis, without writing custom ingest code. Read full definition

Data branches are isolated, full-fidelity copies of a Tinybird workspace, created on demand from main, where schema changes, new pipes, and migrations can be developed and tested against real data before being merged back. Read full definition

Playgrounds are ephemeral SQL scratchpads inside a Tinybird workspace where you can write and run queries against live data sources without saving anything as a pipe or endpoint. Read full definition

Tinybird's SDKs are typed client libraries (Python, TypeScript/JavaScript) that wrap the Events API and published endpoints, so applications can ingest events and read query results without hand-writing HTTP calls. Read full definition