Agents are the New Developer
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.
What Agents Actually Need
Agents work differently than humans. They need:
- MCPs, not UIs – they call functions, not click buttons
- Native search – find the right data instantly
- Fast forks and teardown – spin up experiments without the overhead
- Efficient pricing – pay for what you use
- Built-in knowledge – best practices that come with the database
What We Built
1. An MCP Server That Actually Understands Postgres
We built an MCP server that doesn't just connect to the database—it knows how to use it well. We took 10+ years of Postgres experience and turned it into built-in prompts. Agents get tools for schema design, query optimization, and migrations, plus they can search Postgres docs on the fly.
> 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.
2. Search Built Into the Database
pgvectorscale: We improved our vector search extension. Better indexing throughput, better recall, lower latency than pgvector.
pg_textsearch: Our newest extension. It implements BM25 for proper ranked keyword search, built for hybrid AI apps. Right now it uses an in-memory structure for speed—disk-based segments with compression are coming.
No need to bolt on external search. It's all in Postgres.
3. Instant Database Forks
We built a copy-on-write storage layer that makes databases instantly forkable. Full production data, isolated environment, seconds to create. No data duplication, no cost duplication. You only pay for what changes.
Great for testing, benchmarking, or running migrations in parallel without touching prod.
> Create a fork of my database, test 3 different indexes
> for performance, delete the fork, and report findings.
4. New CLI and a Free Tier
Three commands to get started:
# Install the Tiger CLI and MCP
$ curl -fsSL https://cli.tigerdata.com | sh
$ tiger auth login
$ tiger mcp install
Then either tell your agent to create a free service, or run tiger create service yourself.
Fluid Storage
This all runs on Fluid Storage—our new distributed block store. It's built on local NVMe with a storage proxy that handles copy-on-write volumes.
What you get:
- Instant forks and snapshots
- Automatic scaling, no downtime
- Over 100K IOPS and 1 GB/s per volume
It looks like a local disk to Postgres but scales like cloud storage. Every free service runs on it.
Try It Today
$ curl -fsSL https://cli.tigerdata.com | sh
Built for agents. Designed for developers.
Top comments (1)
This is big