Constructive Open Sources Agentic DB, the Postgres Memory Layer for AI Agents

Built on Constructive's agent-first infrastructure, this open-source Postgres database provides AI agents with persistent memory, structured knowledge, and hybrid retrieval — installable in one command.

SAN FRANCISCO, April 28, 2026 /PRNewswire/ -- Constructive, the company behind open-source Postgres and JavaScript infrastructure with over 100 million open-source downloads, today announced the open-source release of agentic-db — a purpose-built Postgres database architecture that gives AI agents long-term memory, conversation history, a skill and tool registry, task orchestration, runtime observability, and a full CRM and knowledge graph, all inside a single database.

The release arrives as agent frameworks outgrow file-based memory. The prevailing approach — endorsed even by leading AI researchers like Andrej Karpathy — is to store agent knowledge in markdown files. The problem isn't markdown itself; it's that reading markdown files to answer a question is like a library with no Dewey Decimal System — you have to walk every aisle, pull every book off every shelf, and flip through every page just to find a single paragraph. The agent loads its entire history into the context window and hopes the LLM can find what matters before it runs out of space.

agentic-db replaces that pattern. You can still store markdown — but the database searches, filters, and ranks it at the query level. No context window bloat. No token burn. Your agent gets back exactly what it needs.

Kristopher Floyd, Managing Partner at Frontier Syndicate and founder of the Bay Area Frontier Research Club — which convenes researchers from across top-tier labs, including Stanford, UC Berkeley, and Google — worked with Constructive early on to shape agentic-db around a problem he knows firsthand. "I have hundreds of thousands of contacts, and my CRM is a disaster. My agent's only option was to dump everything it knew into the prompt and hope for the best," said Floyd, a former Accenture innovation strategist who advised Fortune 500 leadership on emerging technology. "agentic-db gives agents vector search, full-text search, and auto-embeddings out of the box — so they can actually search what they know instead of brute-forcing it every time. I see Constructive becoming a core piece of infrastructure for the agentic internet."

"We built agentic-db because we believe agent-first infrastructure starts with the database," said Dan Lynch, Founder and CEO of Constructive. "Agents need to remember, search, and reason over structured knowledge — not re-read a pile of markdown files every time they wake up. With Constructive, we're making Postgres the foundation of that infrastructure, and open-sourcing it so every developer building agents can start with a real brain instead of a scratch pad."

What's Inside

agentic-db ships as a single installable Postgres schema, deployable in one command via pgpm — the PostgreSQL Package Manager, which versions, resolves dependencies, and installs database modules the same way npm manages JavaScript packages. The schema covers the full surface area an agent needs:

• Long-term memory — episodic memories with vector, BM25, and spatial search, scoped by agent or shared across multi-agent systems.
• Conversations and tool tracking — messages, tool calls, tool results, and token accounting form a replayable event log of everything the agent has said and done.
• Skills, tools, and prompts — a versioned registry of capabilities with semantic intent matching, so the agent can pick the right skill for a given task.
• Rules and behavioral policies — declarative trigger/action pairs that give agents a governance layer.
• Task orchestration — a priority queue with status, assignment, and result tracking.
• Runtime observability — structured logs, metrics, artifacts, schedules, and an event bus for async workflows.
• World model — a full personal CRM and life-OS: contacts, companies, deals, events, venues, emails, calendars, projects, trips, goals, and habits, connected by 25+ cross-domain junctions.

Every text-bearing table is automatically indexed for five modes of retrieval: semantic vector search via pgvector, BM25 statistical ranking, weighted full-text search, trigram fuzzy matching for typo-tolerant name lookup, and PostGIS spatial queries for geographic reasoning. An auto-embedding pipeline powered by Postgres triggers and Ollama keeps embeddings fresh for semantic search.

Agent-First Infrastructure

Constructive's platform generates Agent Skills and CLIs directly from the database schema — making every database it produces agent-first out of the box. Because agentic-db was built with Constructive, it ships with both automatically.

Agent Skills are structured instruction files that install directly into an agent's workspace and teach it how to use the database. Skills are available for Claude, Claude Code, Cursor, Devin, Copilot, Windsurf, Codex, and 40+ other AI assistants. The project also ships a full CLI with CRUD, search, and admin commands for every table, and a type-safe SDK with a Prisma-like interface. Both are generated from the same database schema and support non-interactive, machine-readable output — purpose-built for agents that need to read and write data programmatically. The same surface lets agentic-db plug directly into frameworks like OpenClaw as a Postgres-backed memory layer.

"Every database Constructive generates is agent-first — the Skills, the CLI, and the SDK all derive from the schema," said Lynch. "agentic-db is the first open-source proof of that. We're giving agents a real database with the tooling already wired up, so developers can focus on what their agents do instead of how they connect."

Michael Shawn Lockwood, a database architect and engineer who has worked at the intersection of SQL Server development and database administration for more than 20 years, deployed agentic-db to a bare-metal Debian VM on day one. "This is one of those rare moments where you stop and think: someone spent years building this," said Lockwood. "What looks like 'just a database' is actually a full foundation layer for agentic systems — long-term memory, conversations, tool execution tracking, task orchestration, hybrid retrieval — and it's all built on top of PostgreSQL."

"I have a strong feeling this becomes a go-to foundation layer for real-world systems over the next few years." — Michael Shawn Lockwood

Local Today, Constructive Cloud Next

agentic-db is released under the MIT license and designed for local and single-developer use — a brain for your agents running on your machine, against your data. For teams and enterprises that need secure Postgres for agents at scale — with multi-tenant deployments, compiled access control, and managed infrastructure — Constructive is building a cloud offering that extends the same schema with the platform's security and provisioning layer.

Availability

agentic-db is available now on npm and GitHub. Developers can get started at pgpm.io. Learn more about Constructive at constructive.io.

About Constructive

Constructive builds agent-first infrastructure, combining modular open-source Postgres with a secure-by-default database architecture designed for the AI era. As agents begin generating and operating applications, Constructive moves trust, access control, and data integrity into the database itself. Its platform spans the entire application stack, from database schemas to APIs to frontend integrations, using a compiler-driven approach to generate and enforce systems end to end. This unified model replaces fragmented tooling with a single, composable foundation for building secure, production-grade applications. Constructive's technology is used by platforms including Supabase, Neon, and Gel Data, and the company has surpassed 100 million open-source downloads. Constructive is led by Dan Lynch, whose previous company Brandcast was backed by Marc Benioff and acquired by TIME. Learn more at constructive.io.

SOURCE Constructive