Introducing Explorations 🗺️
Chat with your data.
Back
May 06, 2025

Explorations: a chat UI for real-time analytics

Introducing Explorations, a conversational UI to chat with your real-time data in Tinybird.
Javier SantanaCo-founder

Analyzing data is time-consuming. It's a manual process of investigating data anomalies, writing SQL, checking SQL reference docs, iterating your SQL, and a lot of fumbling around until you find the issue. I have spent more time in the database console than actually coding, and most of that time was fixing or rewriting queries, reading large text blobs, and comparing numbers. It does not matter what application you are developing; you always need to dig into the data.

Luckily, machines can do that for us now.

Introducing Explorations

While building our own projects, we realized we needed a tool to explore our data, understand its shape, and figure out the right way to query it to get the answers we sought. After a lot of wasted time in the database console, we decided to design what is now called Explorations, a new feature in Tinybird.

Explorations is:

  1. A notebook-style interface to iteratively explore your data through tables and charts.
  2. An AI chat interface to guide your exploration by converting natural language questions into contextualized SQL.
Start building with Tinybird!
If you've read this far, you might want to use Tinybird as your analytics backend. You can just get started, on the free plan.

Exploring code contributions with Explorations

Instead of me writing about all you can do with Explorations, let me show you an actual example:

We store all code contribution events in Tinybird. We use a webhook that sends data from GitLab to Tinybird using our Events API, and then we can analyze that data to get answers to a lot of questions about how we're shipping.. 

GitLab sends a lot of unstructured metadata. It has a bunch of nested JSON, and if you tried to query it manually, you’d have to spend several minutes digging into those large JSON fields to understand how it's all mapped.

Now, we can just ask it: "show me the top 3 contributors”.

We have also published a post on how we built Explorations (feel free to read it), but long story short: we create a custom system prompt for an LLM model that contains all the necessary context about your data (the schemas, data samples, workspace info, etc.), and we give the LLM some tools to run queries, create SQL nodes, etc. It's a very similar experience to our MCP server, but integrated with the UI. 

Subscribe to our newsletter
Get 10 links weekly to the Data and AI articles the Tinybird team is reading.

Start exploring

Explorations has been in beta for several weeks and is now open for everyone. 

If you're new to Tinybird, go install the CLI, create your first workspace, push some data, and start using it. Here's a quick working example so you can test it live:

Go to Explorations and ask questions, for example:

And get a response like this:

Do you like this post? Spread it!

Related posts

Skip the infra work. Ship your first API today.

Read the docs
Tinybird wordmark