Product Demo
Watch what happens when a human talks to an agent and the agent talks to Chaprola
We recorded what it looks like when an AI agent uses Chaprola from scratch. No setup in advance. No staging environment. No demo magic. A human types five messages in plain English. The agent does everything else.
What happens in the video
1. Install. The human says "Add Chaprola to my setup." The agent installs the MCP server, registers an account, and signs the HIPAA Business Associate Agreement. Two lines from the human. Three API calls from the agent.
2. Import. The human says "Import the NYC yellow taxi trip data from January 2024." The agent fetches over a million records directly from the NYC Taxi and Limousine Commission. Eighty-eight megabytes, imported in eight seconds. No CREATE TABLE. No manual schema.
3. Build a report. The human describes what they want: total trips, total fares, averages, highest and lowest fare. The agent writes a Chaprola program, compiles it to bytecode, and runs it against a million records in 7.59 seconds.
4. Publish. The human says "Publish that so anyone can view it." The agent publishes the report to a public URL. Anyone with the link can run it -- no login, no API key. The program executes live against the real data every time.
5. Make it pretty. The human says "Now make it pretty." The agent reformats the report however you like.
The result is a live report, running right now, processing a million records on demand:
Why this matters
Every tool in this demo is a REST endpoint. Every step the agent took is an HTTP call. There is no database to provision, no driver to install, no infrastructure to manage. The agent operated autonomously from start to finish.
This is what an agent-first data platform looks like. The human decides what they want. The agent figures out how to get it done. Chaprola handles the rest.
The numbers
- 5 messages from the human
- 8 API calls from the agent
- 1,050,000 records imported
- 7.59 seconds to process
- 1 live URL, no expiration
The video shows both sides -- what the human types on the left, what the agent executes on the right. Watch both sides.