TLDR: TigerData, formerly Timescale, has launched ‘Agentic Postgres,’ the first database specifically designed for AI agents. This new solution addresses the unique needs of autonomous AI systems with features like instant, isolated database forks for safe experimentation, native full-text and semantic search, and a Master Control Program (MCP) server that helps agents interact intelligently with the database. The company also announced a free plan to foster AI-driven development.
In a significant move for the artificial intelligence and database sectors, TigerData, previously known as Timescale, announced today, October 21, 2025, the launch of ‘Agentic Postgres.’ This innovative database is positioned as the first of its kind, purpose-built to support the burgeoning era of AI agents and their distinct operational requirements. The announcement underscores a fundamental shift in software development, where AI agents are increasingly becoming integral to the coding and operational landscape.
According to Ajay Kulkarni, co-founder and CEO of TigerData, the rise of AI agents necessitates a re-evaluation of traditional data infrastructure. “Modern AI depends on live context, but traditional databases weren’t designed for agents that act,” Kulkarni stated. “You need infrastructure that lets them branch, test and learn without touching production.” He highlighted the rapid acceleration of AI in code generation, noting that “80% of Claude Code was written by AI” and that “more than a quarter of all new code at Google was generated by AI one year ago.” This trend suggests that the majority of new code will soon be AI-generated, demanding tools that cater to how agents operate: they ‘call’ instead of ‘click’ and ‘retrieve’ instead of ‘remember.’
Agentic Postgres is engineered with several core features to meet these demands:
Fluid Storage with Instant Forks: A cornerstone of Agentic Postgres is its ability to create instant, isolated database copies, referred to as ‘forks.’ This allows AI agents to test, simulate, and experiment with new schemas or retrieval strategies without any risk to live production data. This capability facilitates hundreds of parallel experiments safely and cost-effectively, a critical need for iterative AI development.
Native Search and Retrieval: The database integrates advanced search functionalities directly. For semantic search, it leverages an improved `pgvectorscale` extension, offering higher-throughput indexing, better recall, and lower latency at scale. For full-text search, the new `pg_textsearch` extension implements BM25, optimized for hybrid AI applications and designed for fast writes and queries with an in-memory structure in its preview release.
Master Control Program (MCP) Server: Agentic Postgres includes a novel MCP server that empowers agents not just to interact with the database, but to understand and utilize it effectively. This server distills TigerData’s decade of Postgres expertise into built-in master prompts, providing agents with structured access to high-level tools for schema design, query tuning, and migrations. It also performs native full-text and semantic search over Postgres documentation, enabling agents to retrieve relevant context instantly.
Emphasis on Safety and Transparency: Recognizing the autonomy of AI agents, TigerData has prioritized safety at the data layer. Every fork created by an agent leaves an immutable data trail, complete with version history and memory state. This ensures transparency, allowing engineers to inspect, replicate, or undo any autonomous action, keeping humans firmly in the loop.
TigerData emphasizes that Agentic Postgres is designed to elevate developers, not replace them. “Agents are not here to replace us. They’re here to elevate us,” the company stated. By automating mechanical and repetitive tasks, agents free developers to concentrate on architecture, design, creativity, and impact. Kulkarni added, “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.”
To further support this new paradigm, TigerData also announced the Tiger Free Plan. This fully managed Postgres offering provides free access with no credit card, no time limit, and no hidden catches, specifically tailored for prototyping and experimentation in AI development. This initiative aims to lower the barrier for developers and agents to build together on Postgres without friction.
Also Read:
- Aisera Advances Autonomous IT with Third-Generation Agentic AI Platform for Enhanced Enterprise Self-Service
- Amplitude Unveils Model Context Protocol and Expands AI Agents for Automated Behavioral Data Analysis
TigerData, backed by $180 million from top-tier investors, has a history of scaling PostgreSQL across millions of workloads and serving thousands of customers worldwide, including major players like Mistral, HuggingFace, Nvidia, Toyota, Tesla, NASA, and JP Morgan Chase. The launch of Agentic Postgres marks a new chapter in their mission to provide modern PostgreSQL solutions for the analytical and agentic era.
