@tinybird
Account based inSpain
About this account
- Account based in
- Spain
- Connected via
- Web
Account-level information from X, not a live location or the device used for a specific post.
Tinybird is a managed ClickHouse® service for AI-native software teams. Get ClickHouse performance without ClickHouse complexity.
Joined April 2019
- Tweets4.7K
- Following310
- Followers8.5K
- Likes3.3K
Classic was Tinybird’s original experience. Forward is the next big iteration, built for teams operating data products: deployments, schema migrations, local dev, CI/CD, SDKs and Agent Skills.
Free and Developer Classic workspaces must migrate by Sep 15: tinybird.co/blog/tinybird-cl…
Use separate compute for CDC snapshots, deduplication, and rollups.
Scheduled: `ON_DEMAND_COMPUTE true`
Manual: `tb copy run my_copy_pipe --on-demand-compute`
Provisioning adds overhead and compute is billed.
tinybird.co/blog/on-demand-c…
Unpredictable event payloads usually force a bad tradeoff: rigid upfront migrations, or raw string columns that waste CPU parsing JSON on every single query.
Tinybird's native JSON data type is now on by default across all workspaces.
ALT Tinybird pipe node extracting typed fields from a native JSON column using dot notation and type casts
Store nested payloads in ClickHouse with automatic type inference.
Query subcolumns with dot notation and casts like payload.user_id::String. No schema shredding required.
tinybird.co/docs/changelog/2…
Deduplication happens during asynchronous merges, so duplicate versions can remain temporarily. Use FINAL when a read needs the deduplicated state now, after considering its query-time cost.
tinybird.co/docs/forward/gui…
Tinybird retweeted
I'm building a long-horizon agent today, and Tinybird (@tinybird) gives the history a queryable home: the agent asks about its own past instead of carrying all of it in context.
-
@jorgesancha
nitter.cf/AgentGlassInc/status/2…
At the AWS Builder Loft (@AWSstartups) in San Francisco, Tinybird (@tinybird) said it forked ClickHouse to make it schemaless. Send any raw JSON event, query it with SQL, never write a migration. It is called RawTree, and it launches in the coming weeks.
Tinybird has run managed ClickHouse since 2019. Enzo Kajiya, who leads US sales, is testing a tagline: "If MongoDB and ClickHouse had a baby, it would look a lot like RawTree."
Built agent-native from the start: an MCP server, skills, a CLI. OTel-native.
Three use cases: observability, real-time analytics, and what he called sandbox and agent observability. An agent can query the data, and it can investigate its own work.
Also auto-scaling, bring-your-own-cloud and bring-your-own-bucket.
The hackers in this room get to build on it before launch.
@SidraMiconi · @AgentGlassInc
-
@tokensandai · @inaccisland · @ale_amenta · @jorgesancha · @bnevilleoneill . @awscloud . @AWS
Tinybird retweeted
At the AWS Builder Loft (@AWSstartups) in San Francisco, Tinybird (@tinybird) said it forked ClickHouse to make it schemaless. Send any raw JSON event, query it with SQL, never write a migration. It is called RawTree, and it launches in the coming weeks.
Tinybird has run managed ClickHouse since 2019. Enzo Kajiya, who leads US sales, is testing a tagline: "If MongoDB and ClickHouse had a baby, it would look a lot like RawTree."
Built agent-native from the start: an MCP server, skills, a CLI. OTel-native.
Three use cases: observability, real-time analytics, and what he called sandbox and agent observability. An agent can query the data, and it can investigate its own work.
Also auto-scaling, bring-your-own-cloud and bring-your-own-bucket.
The hackers in this room get to build on it before launch.
@SidraMiconi · @AgentGlassInc
-
@tokensandai · @inaccisland · @ale_amenta · @jorgesancha · @bnevilleoneill . @awscloud . @AWS
Sunday read: Materialized Views process incoming blocks incrementally. When a transformation needs broader data, a scheduled Copy Pipe can build the batch view. That work can now run on temporary dedicated compute.
tinybird.co/blog/on-demand-c…
Scheduled Copy Pipes can now run on temporary dedicated compute. Add `ON_DEMAND_COMPUTE true` to keep recurring batch work separate from primary workspace compute.
How it works and when provisioning is worth it:
tinybird.co/blog/on-demand-c…
ALT Tinybird Copy Pipe configured to run on a schedule using on-demand compute
Building microservices just to expose ClickHouse queries to your app creates unnecessary plumbing.
In Tinybird, write parameterized SQL with template variables like {{ Int32(hours, 24) }}. Tinybird publishes an authenticated, production REST API instantly.
ALT Tinybird SQL pipe endpoint with dynamic query parameters and quantile latency calculation
Published endpoints include typed parameters, token authentication, caching, and service data source logs.
Test with tb endpoint data, then deploy.
tinybird.co/docs/forward/cor…
Test ClickHouse schema changes without touching production:
1. Create an isolated branch: tb branch create
2. Validate the change
3. Deploy: tb deploy
Tinybird applies safe changes without downtime.
tinybird.co/docs/forward/cor…
Scanning raw events for every dashboard view wastes compute and adds latency.
Tinybird Materialized Views aggregate streaming data on insert. Define a Pipe with countState() or uniqState(), and rows roll up into AggregatingMergeTree tables continuously.
ALT Tinybird Materialized Pipe aggregating streaming events into an AggregatingMergeTree data source
Query the rollup with countMerge() and uniqMerge(). Endpoints read pre-aggregated states instead of scanning raw events across the full time range.
tinybird.co/docs/forward/cor…
We’re at Rows & Columns Summit today 👋
Visit our booth to talk real-time data with us and get an early look at @rawtreedb, what we’re building next.
Scan the QR code and leave your details to enter our iPad raffle.
📍 Contemporary Jewish Museum
Streaming data into ClickHouse often starts with standing up Kafka, connectors, and consumer fleets.
With Tinybird's Events API, you POST JSON directly over HTTP at 1K+ events/sec. Data lands in MergeTree tables and is queryable in seconds.
ALT cURL request sending a JSON page view event to the Tinybird Events API
Microbatch with NDJSON for high throughput. When you need guaranteed persistence before continuing, pass wait=true for immediate HTTP 200 write acknowledgements.
No pipeline plumbing required.
tinybird.co/docs/forward/ing…
ClickHouse is fast. Running it means managing replicas, Keeper, upgrades, and part merges.
Tinybird gives you managed ClickHouse with a 99.9% uptime SLA, plus ingestion and APIs.
tinybird.co/product/managed-…
With Jev with realize that when people get something that's faster (even with some limitations) they find new ways and new use cases.
I saw this with data, when people started to use ClickHouse instead of Spark (or even Snowflake) they naturally used it for new use cases and solve problems they didn't even consider before.
A good lesson to learn, speed is more than a feature.