Why Agents Love Supabase — with Paul Copplestone and Ant Wilson
Summary
Supabase co-founders Paul Copplestone and Ant Wilson detail the company's strategy for becoming the preeminent open-source, integrated backend for AI-built applications. The core focus is on enabling AI agents by providing a seamless 'build, operate, and scale' experience. Key technical advancements include collocated compute (Supabase Compute), performance enhancements (OrioleDB), and massive scaling solutions (Multigress and the acquisition of Turso) designed to handle billions of databases per month.
Key takeaways
-
Agent-Centric Development Model
38:35
The platform is evolving to support AI agents by prioritizing a 'code-first' approach. This includes using declarative schemas (via PG Delta) to allow agents to understand the full shape of the database (tables, policies, comments) directly from the file system, rather than relying on traditional migration files.
-
Integrated Compute and Database
38:35
Supabase Compute offers a long-running, serverless, collocated compute environment (similar to a containerized Lambda function) running on the same private network as the database. This minimizes latency and allows agents to run complex workloads (like chunking and embedding) directly from the database connection.
-
Scaling to Billions of Databases
20:50
To handle massive scale, Supabase announced Multigress, a groundup rewrite of the Postgres operator designed for sharding. The goal is to allow zero-code changes for users scaling their databases, alongside the acquisition of Turso, which specializes in MVCC for SQLite, enabling agentic databases.
-
Product-Led Growth (PLG) Strategy
1:09:03
The company's enterprise strategy focuses on winning 'green field' use cases (new builds) rather than complex database migrations. They leverage product triggers (e.g., hitting scaling limits or errors) to initiate sales conversations, making the platform sticky from day one.
Technical details
-
Database Performance & Reliability
2315s
OrioleDB is a drop-in replacement for the current Postgres `heap` storage engine, offering 1.5 to 2 times faster performance and removing the need for background tasks like `vacuum` cleanup. It is fully compatible with Postgres and released under an Apache 2 license.
-
Database Scaling (Sharding)
2315s
Multigress is a major rewrite of the Postgres operator designed to manage multi-node clusters and sharding. It aims to provide sharding capabilities with zero required code changes for the user, ensuring high stability even after rigorous testing (1,000 sequential failovers).
-
Local Development & Schema Management
2315s
The introduction of declarative schemas, facilitated by the PG Delta library, allows developers to represent the full shape of a Postgres database (tables, policies, comments) in a file system folder. This is critical for agents, as it provides a complete, easily crawlable source of truth without needing to run slow introspection queries.
-
Compute Architecture
2315s
Supabase Compute provides a full, long-running, serverless compute instance collocated with the database. Unlike previous limitations (e.g., Deno's 2C limit), this environment supports multiple languages (Rust, Python) and is designed to host complex agentic workloads.
Mentioned resources
Channel & topics
Watch on YouTube · Back to latest
This independent, AI-assisted summary is provided for commentary and informational purposes. It may contain errors or omit important context. Please watch the original video for the creator's complete presentation. Video, thumbnail, and related copyrights belong to their respective owners.