Tinybird

A hosted ClickHouse platform that turns SQL queries into authenticated REST API endpoints, for analytics inside your product.

What it is: Real-time analytics databases
Type: Commercial product
Made by: Tinybird
License: Proprietary (TypeScript SDK MIT-licensed)
Facts checked: 6 October 2026

What it is

In Tinybird you define Data Sources, which are ClickHouse tables, and Pipes, which are chains of SQL queries called nodes. A Pipe marked TYPE ENDPOINT becomes an HTTP API at https://api.tinybird.co/v0/pipes/<name>.json (or .csv, .parquet and others), with typed query parameters filled in by a template language and access controlled by static tokens or short-lived JWTs that a browser can use per tenant. Data comes in through the Events API (JSON over HTTP), the Kafka connector, S3 and GCS connectors, a DynamoDB connector that follows table changes, and table functions that copy from PostgreSQL, MySQL, Iceberg or URLs on a schedule. Streaming writes pass through a batching service before they reach ClickHouse, all data is stored in S3 or GCS, and each ClickHouse replica keeps a local SSD cache; replicas share the stored data through zero-copy replication, which Tinybird keeps alive in its own ClickHouse fork. Materialized views pre-aggregate on every insert, and BI tools can connect through a read-only ClickHouse HTTP interface.

The current product is Tinybird Forward, launched in March 2025. A project is a folder of plain-text datafiles (or TypeScript or Python SDK code) that you build against Tinybird Local, a Docker container, or a cloud branch, then deploy; the deployment works out the migrations and backfills a schema change needs. The older Tinybird Classic is being retired: Free and Developer workspaces had to move to Forward by 15 September 2026, and those left are scheduled for automatic migration. Tinybird runs in AWS and Google Cloud regions. Free gives 0.25 vCPU, 1,000 requests a day and 10 GB. Developer plans cost $25 to $799 a month for 0.25 to 8 vCPUs on shared infrastructure, with 25 GB included and then $0.058 per GB; use above the plan’s vCPUs is billed at $0.0002 per vCPU-second. SaaS (shared, up to 32 vCPUs, 12-month commitment) and Enterprise (a dedicated ClickHouse cluster) are priced by sales (list prices, checked 6 October 2026).

Why teams pick it

From a SQL query to a parameterised, token-protected HTTP API without writing a backend; JWTs let browsers query per tenant
ClickHouse performance without running clusters, Keeper or storage tiering yourself
Projects are files in Git: Tinybird Local in Docker for development and CI, cloud branches for previews, and deployments that plan backfills for schema changes
Managed ingestion from Kafka, S3, GCS and DynamoDB, plus sinks back to S3, GCS and Kafka
A free plan with no time limit, and self-service plans you can resize at any time

Watch out for

Proprietary and hosted only, in AWS and Google Cloud regions; self-managed regions are in beta and meant for development only
Endpoints, tokens and datafiles are Tinybird-specific: the SQL is ClickHouse SQL, but leaving means rebuilding the API and ingestion layers
Shared plans cap compute: requests beyond the burst limit get HTTP 429, extra vCPU time is billed per second, and queries stop after 10 to 30 seconds depending on the plan
It runs Tinybird’s ClickHouse fork, so upstream features arrive on Tinybird’s schedule, and the ClickHouse interface is read-only
Table functions only reach databases exposed to the internet with publicly signed certificates

Alternatives

What all real-time analytics databases have in common →

Alibaba Cloud HologresCommercialAlibaba Cloud’s PostgreSQL-compatible real-time data warehouse, which takes streaming writes, answers analytical queries and serves key lookups from one copy of the data.

Apache DorisOpen sourceAn open-source MPP analytical database, born at Baidu as Palo, that serves sub-second queries on fresh, updatable data through the MySQL protocol and also queries Iceberg, Hudi, Paimon and Hive tables in place.

Apache DruidOpen sourceAn open-source real-time analytics database that makes Kafka and Kinesis events queryable within seconds and serves fast slice-and-dice queries to dashboards and APIs.

FireboltCommercialA cloud data warehouse built for low-latency, high-concurrency queries behind apps and dashboards, also free to self-host as Firebolt Core.

PivotOpen sourceA young open-source analytics server in Rust that speaks the PostgreSQL protocol and queries Iceberg and Delta Lake tables straight from object storage.

StarRocksOpen sourceAn open-source MPP analytical database with a vectorized engine that serves fast joins and real-time updates on its own tables, and also queries Iceberg, Hudi, Delta Lake and Hive tables in place.

How it fits with other technologies

What Tinybird uses

Build a Data Source and an Endpoint in Tinybird Local, call it, then deploy · bash

curl https://tinybird.co | sh          # installs the tb CLI
tb init --type cli --dev-mode local --folder .
tb local start --daemon                # Tinybird Local: a Docker container on :7181

mkdir -p datasources endpoints
cat > datasources/page_views.datasource <<'EOF'
SCHEMA >
    `timestamp` DateTime `json:$.timestamp`,
    `pathname` String `json:$.pathname`,
    `session_id` String `json:$.session_id`

ENGINE "MergeTree"
ENGINE_SORTING_KEY "pathname, timestamp"
EOF

cat > endpoints/top_pages.pipe <<'EOF'
TOKEN dashboard READ

NODE aggregated
SQL >
    %
    SELECT pathname, count() AS views
    FROM page_views
    WHERE timestamp >= {{DateTime(start_date)}}
    GROUP BY pathname
    ORDER BY views DESC
    LIMIT {{Int32(limit, 10)}}

TYPE ENDPOINT
EOF

tb build                               # creates the table and the endpoint locally
TOKEN=<admin token from: tb token ls>

curl -X POST "http://localhost:7181/v0/events?name=page_views" \
  -H "Authorization: Bearer $TOKEN" \
  -d '{"timestamp":"2026-10-06 10:00:00","pathname":"/pricing","session_id":"s1"}'

curl -H "Authorization: Bearer $TOKEN" \
  "http://localhost:7181/v0/pipes/top_pages.json?start_date=2026-10-06%2000:00:00"

tb --cloud deploy                      # same files, deployed to Tinybird Cloud

Plans#

List prices from tinybird.co/pricing and the plan docs, checked on 6 October 2026. Every plan includes unlimited seats, pipes and endpoints.

	Free	Developer	SaaS	Enterprise
Price	$0	$25 to $799 a month (0.25 to 8 vCPUs)	Custom, 12-month commitment	Custom
Infrastructure	Shared	Shared	Shared, up to 32 vCPUs	Your own ClickHouse cluster
Included storage	10 GB	25 GB, then $0.058 per GB	From 500 GB, then $0.058 per GB	Bottomless, $0.058 per GB
Over the plan’s vCPUs	Hard limit, HTTP 429 past 3× burst	$0.0002 per vCPU-second	$0.0002 per vCPU-second	Not applicable
Requests	1,000 a day	Unlimited	Unlimited	Unlimited
Query time limit	10 s	20 s	20 s	30 s

Egress through sinks costs $0.01 per GB within the same cloud and region and $0.10 per GB across clouds or regions. Large backfills and some copy jobs run on on-demand CPUs, billed in credits per core-minute by region.
Before January 2025, Tinybird billed by data processed; it now bills by vCPU.

See it in a real stack

No architecture uses Tinybird yet. Drop it into the playground and wire it up yourself.Open playground →

Benchmarks

Benchmark · ClickHouse Inc.ClickBenchHow fast analytical databases run 43 SQL queries over one 100-million-row table of web analytics events, on the same cloud machines.Includes Tinybird · 7 caveats to read first

How to read a benchmark →

Where it came from2019–2026

Tinybird’s founders say it began with their own problem: their companies had plenty of data, but it sat in a cloud data warehouse behind gatekeepers, and the developers building products couldn’t reach it quickly. They set out to turn data at any scale into a real-time, secure REST API, and built the service on ClickHouse. By September 2019 the Madrid team was demoing “Tinybird Analytics” at South Summit, showing sub-second queries over billions of rows.

In April 2022 co-founder Jorge Sancha announced a $37 million Series A from CRV, Singular and Crane, and in June 2024 a $30 million Series B led by Balderton Capital. In January 2025 pricing moved from data processed to vCPUs. In March 2025 Tinybird launched Forward, rebuilt around projects as files, local development and deployments, after large customers said schema changes and promotion between environments had become painful. Classic was retired for Free and Developer workspaces in 2026.

Why it’s built this way

Publish queries as HTTP endpoints, not database connections: The users were application developers. An endpoint with typed parameters and scoped tokens, later JWTs, can be called from a backend or a browser without writing an API service in front of the database.
Build on ClickHouse, then keep a fork: ClickHouse gave the speed. Tinybird relies on zero-copy replication over S3 and GCS, which upstream ClickHouse said it would remove, so in 2025 it explained why it maintains its own fork with that feature and its own release process.
Serverless plans instead of picking cluster sizes: Sancha wrote in 2022 that users wanted fast, reliable queries and not to choose CPU cores and memory per use case, so Tinybird sells plan sizes on shared infrastructure and handles the ClickHouse tuning.
Projects as code with planned deployments: In Classic a schema change could mean creating a new Data Source, copying data and switching consumers by hand. Forward analyses the dependency graph and plans the backfills and table swaps itself.

Timeline

2019
Demos Tinybird Analytics at South Summit Madrid in September
2022
Raises a $37 million Series A
2024
Raises a $30 million Series B led by Balderton Capital
2025
Moves to vCPU-based pricing; launches Tinybird Forward in March
2026
Retires Classic for Free and Developer workspaces in September

Sources24 · facts checked 6 October 2026

Improve this page on GitHub