Tinybird
Real-time analytics platform for building data APIs over ClickHouse
Tinybird makes Explorations shareable and tightens viewer and guest permissions.
◆Recent moves
- 6d ago
Share Explorations with your Workspace
Shared Explorations, a Workspace filter on org observability and tighter viewer and guest permissions push Tinybird Forward toward team use. It follows the Classic sunset that moved smaller customers onto Forward.
View source ↗ - 13d ago
Tinybird Classic sunset for Free and Developer plans is today
The scheduled Classic sunset takes effect for Free and Developer plans, and CLI 5.0.0 raises the minimum to Python 3.11. Smaller users now have only Forward, which finishes much of the migration the recent releases were built around.
View source ↗ - 20d ago
Remove several organization members at once
Admins can now remove organization members in bulk from the UI or API, dedicated clusters gain protection logs, and failed Sink jobs appear in sinks_ops_log. After weeks of query and ingest tuning, this release shifts to the admin and observability needs of larger accounts.
View source ↗ - 27d ago
Tinybird Local on arm64 requires a 4KiB page size kernel
Secondary indexes and a configurable deduplication window give users direct control over query speed and ingest semantics, two things they previously had to accept as defaults. The 4KiB page-size requirement for arm64 Local is a real compatibility catch for developers on some ARM Linux hosts, and stricter validation fits the post-Classic cleanup.
View source ↗ - 1mo ago
Run scheduled Copy Pipes on on-demand compute
On-demand compute for scheduled Copy Pipes eliminates the need to pre-provision capacity for batch-style data movement jobs. The single MCP tool for Endpoint calls simplifies the tooling surface for LLM integrations, and job-based remote file imports extend the ingestion API to pull from external URLs directly — a useful complement to the push-based model.
View source ↗ - 1mo ago
Choose MCP response formats and inspect query plans
Configurable MCP output formats give LLM integrations control over response shape, and query plan inspection surfaces execution details that are otherwise opaque. Together they make the MCP tooling more production-ready for LLM workflows that need to tune query behavior rather than accept default outputs.
View source ↗