Tinybird lets you turn any SQL query into a secure, low-latency REST API. Define your query with dynamic parameters, publish it, and consumers call it as a standard HTTP endpoint. Responses are in JSON by default, with CSV and NDJSON also available. Read the Endpoints docs.

Any tool that supports the ClickHouse protocol: Grafana, Tableau, Metabase, Looker Studio, Power BI, Superset, Hex, DataGrip, DBeaver, and any other ClickHouse-compatible client. Connection is over HTTPS on port 443. Read the ClickHouse interface docs.

A hosted Model Context Protocol server that gives AI agents structured access to your Tinybird workspace. Agents can explore data sources, run SQL queries, list and call your API endpoints. Every published endpoint is also exposed as an MCP tool that agents can discover and call. Authentication uses standard Tinybird tokens with enterprise-grade isolation. Read the MCP Server blog post.

All interfaces use Tinybird tokens. You can create static tokens or use JWTs for dynamic, user-scoped access. Row-Based Access Control (RBAC) policies let you control exactly which data each consumer — whether an application user, BI tool, or AI agent — can see. Learn about token permissions.

API Endpoints respond in milliseconds, backed by ClickHouse's columnar engine. Tinybird optimizes query execution and caching automatically. Latency depends on query complexity and data volume, but sub-100ms responses are typical for well-structured queries.

Yes. Any client that speaks the ClickHouse HTTP protocol can connect: DataGrip, DBeaver, clickhouse-client, or custom scripts using the ClickHouse HTTP API. Connection is read-only by design.

All queries are tracked in service data sources. BI tool queries appear in bi_stats_rt and bi_stats. MCP and API queries appear in pipe_stats_rt and pipe_stats. You can query these tables for free from Playgrounds, Time Series, or your own monitoring.

The MCP Server exposes: explore_data (inspect data sources), text_to_sql (natural language to SQL), execute_query (run SQL), list_endpoints and list_datasources (discover resources). Every published API endpoint is also exposed as an MCP tool that agents can discover and call.

Yes. Every endpoint gets auto-generated OpenAPI documentation. You can test parameters and preview responses directly from the docs page. The API host and endpoint URLs are accessible from the workspace overview.

SQL APIs (Endpoints) are purpose-built REST APIs with dynamic parameters, caching, and authentication, designed for applications and dashboards. The ClickHouse interface exposes your data sources for ad-hoc SQL queries from BI tools and SQL clients. Use SQL APIs when you need a productized, optimized endpoint. Use the ClickHouse interface when you need flexible, exploratory access.

Yes. The Tinybird CLI provides three query commands: tb endpoint data to call any published endpoint with parameters, tb sql to run arbitrary SQL over your pipes and data sources, and tb ch to query directly via the ClickHouse HTTP interface. Read the CLI docs.

Yes. The TypeScript SDK and Python SDK let you define and call endpoints programmatically. They are primarily for building your data project, but you can also use them to query endpoints from application code.

Materialized Views let you pre-aggregate and transform data at ingest time. Instead of running expensive aggregations on every query, Tinybird continuously updates a derived dataset as new data arrives. This reduces query latency and cost per query, especially for high-volume use cases like counters, averages, and top-N rankings. Read the Materialized Views docs.

Yes. Tinybird Sinks let you push processed data to Kafka topics, S3 buckets, or Google Cloud Storage. You can schedule exports with cron expressions or trigger them on-demand, and choose between CSV, NDJSON, and Parquet formats. Sinks are fully managed, so there is no infrastructure to maintain.

Yes. Tinybird Explorations is an AI-powered interface where you ask questions in natural language and get results without writing SQL. For hands-on SQL work, Playgrounds provide an interactive editor for prototyping queries before publishing them as APIs. Time Series lets you visually explore time-based data with charts and drill-down. Learn more on the Workspace page.