Category: All posts
Oct 21, 2025
80% of Claude Code was written by AI. More than a quarter of all new code at Google was generated by AI one year ago. It’s safe to say that in the next 12 months, the majority of all new code will be written by AI.
Agents don’t behave like humans. They behave in new ways. Software development tools need to evolve. Agents need a new kind of database made for how they work.
But what would a database for agents look like?
At Tiger, we’ve obsessed over databases for the past 10 years. We’ve built high-performance systems for time-series data, scaled Postgres across millions of workloads, and served thousands of customers and hundreds of thousands of developers around the world.
So when agents arrived, we felt it immediately. In our bones. This new era of computing would need its own kind of data infrastructure. One that still delivered power without complexity, but built for a new type of user.
How do agents behave?
We started on this problem over a year ago. Multiple teams working in parallel, months of engineering and internal user feedback, rethinking everything from the storage layer to how agents actually reason.
Here’s what we built.
Today we’re launching Agentic Postgres, the first database designed from the ground up for agents. It includes:
The best database MCP server ever built
Agentic Postgres includes our new MCP server that enables agents not just to interact with the database, but also understand how to use it well. We’ve taken our 10+ years of Postgres experience and distilled it into a set of built-in master prompts. This gives agents safe, structured access to the database through high-level tools for schema design, query tuning, migrations, and more. The MCP server also performs native full-text and semantic search over the Postgres docs, so agents can instantly retrieve the right context as they think.
> I want to create a personal assistant app. Please create a free service on Tiger. Then using Postgres best practices, describe the schema you would create.
Native search and retrieval
Agentic Postgres comes with native full-text and semantic search built directly into the database. For semantic search, we’ve improved our existing extension pgvectorscale, for higher-throughput indexing, better recall, and lower latency at scale than pgvector.
For full-text search, pg_textsearch, our newest Postgres extension, implements BM25 for modern ranked keyword search optimized for hybrid AI applications alongside pgvectorscale. The current preview release uses an in-memory structure for fast writes and queries. Future releases will add disk-based segments with compression and BlockMax WAND optimization, applying the same battle-tested techniques from production search engines.
Together, these extensions let agents retrieve structured data instantly without leaving Postgres.
> Using service qilk2gqjuz, analyze user feedback with hybrid search (combining text search and semantic search). Group similar feedback by theme and show counts for each theme, using an ascii bar chart. First, look at the pg_textsearch (BM25) and pgvectorscale documentation in the Tiger docs to get the proper syntax, and then use those extensions.
Fast, zero-copy forks
At the core of Agentic Postgres is a new copy-on-write block storage layer that makes databases instantly forkable. Every agent can spin up its own isolated environment, a full copy of production data in seconds, without duplicating data (or costs). Every fork is lightweight and efficient, so you only pay for the blocks that change. It’s perfect for experiments, benchmarks, and migrations that can run safely in parallel.
> Please create a fork of gf868h9j1y using the last snapshot, and then test 3 different indexes that we should create to speed up performance, then delete the fork, and report back on your findings. Before you start run “tiger service fork --help” and “tiger service delete --help” to get the right syntax. Use MCP over psql, using the password from the local keychain.
New CLI and free tier
We’ve also built a new CLI that makes it easy to explore, fork, and build with Agentic Postgres, and we’re launching a free tier so every developer and every agent can get hands-on right away.
This is all launching today. You can try this today with 3 basic commands in your terminal:
# 3 commands to install the Tiger CLI and MCP. That's it!
$ curl -fsSL https://cli.tigerdata.com | sh
$ tiger auth login
$ tiger mcp install
Then just tell your agent to spin up a new free service using MCP, or simply call tiger create service
from the command line to get going.
Agentic Postgres is powered by Fluid Storage, our new distributed storage layer. Fluid Storage is built on a disaggregated architecture of a horizontally scalable distributed block store using local NVMe storage, a storage proxy layer that exposes copy-on-write volumes, and a user-space block device driver.
It’s storage that looks like a local disk to Postgres, yet scales like a cloud service.
As a result, Fluid Storage delivers instant forks, snapshots, and automatic scaling (up or down) without downtime or over-provisioning. In benchmark testing, a single volume sustains throughput over 100,000 IOPS, while retaining all of Fluid’s elasticity and copy-on-write capabilities.
All free services on Tiger Cloud run on Fluid Storage today, so every developer can experience its performance and flexibility firsthand.
And this is just the start. We’ll dive deeper into each of these (MCP, pg_textsearch, pgvectorscale, forkable databases, Fluid Storage, CLI, free tier) later this week and next.
Agentic Postgres is built for agents, so developers can work on higher level problems.
Building with and for agents, we’ve learned something simple: agents are not here to replace us. They’re here to elevate us.
Agents take on the mechanical, repetitive work, freeing us to focus on what matters most: architecture, design, creativity, impact. They make us faster, smarter, and enable us to do more ambitious work than we could do alone.
The myth is that AI will replace developers. The truth is that developers who build with agents will replace those who don’t.
Agentic Postgres is for developers who want to build with AI. For developers who care more about working applications than disposable demos. For developers who want AI to feel like engineering, not just experimentation.
Today’s launch is just the beginning. There are still some rough edges. We’d love your help sanding them down. But expect more to come: more launches, big and small, in the weeks, months, and years ahead.
Agents are the new developers. Agentic Postgres is their new playground.
Built for Agents. Designed to Elevate Developers.
So let’s build. Together.
Get started today:
$ curl -fsSL https://cli.tigerdata.com | sh