vector databases 100% Deterministic Spec Audit

pgvector vs Supabase Vector

Comparing architecture, runtime limits, open-source licenses, and monthly costs for autonomous agent pipelines.

pgvector

Andrew Kane & PostgreSQL Community

Supabase Vector

Supabase Inc. (Paul Copplestone & Ant Wilson)

Detailed Specifications & Compatibility Matrix

Extracted directly from GitHub REST API & official developer documentation

Technical Feature / Metric pgvector Supabase Vector
Developer / Organization Andrew Kane & PostgreSQL Community Supabase Inc. (Paul Copplestone & Ant Wilson)
Open-Source Engine ✓ Open Source (Public) ✓ Open Source (Apache-2.0)
GitHub Stars (Repo/SDK) ★ 23,475 GitHub Stars ★ 113,917 GitHub Stars
Core Repository Language C TypeScript
Supported SDK Clients
sqlpythontypescriptrustgo
typescriptpythondartswiftgo
Self-Hosted Deployment ✓ Supported ✓ Supported
Docker / Container Ready ✓ Official Image ✓ Official Image
Model Context Protocol (MCP) ✓ Native MCP ✓ Native MCP
Billing Architecture free hybrid
Developer Free Tier
✓ Yes

100% Free (PostgreSQL license)

✓ Yes

500MB database storage (2 free projects)

Entry Paid Tier (Monthly) Free / Custom $25 USD / mo
Pricing & Cost Architecture 100% Free and open-source under the PostgreSQL License. Pre-installed and available with zero license fees on Supabase, AWS RDS, Neon, Azure, and self-hosted Postgres. Free tier offers 2 managed projects with 500MB database storage (~250k vectors). Pro plan starts at $25/mo with 8GB storage and scalable compute add-ons.
Key Engineering Insight Changed the vector landscape by demonstrating that 90% of RAG architectures do not need a standalone vector database; they can simply execute vector queries on their existing Postgres tables. Allows developers to store relational application state, user authentication (Row Level Security), and high-dimensional vector embeddings inside a single ACID-compliant Postgres database.

Strengths & Trade-Offs

Objective engineering evaluation of both platforms

pgvector Evaluation

Key Advantages

  • ✓ Runs directly inside PostgreSQL: eliminates data synchronization pipelines, extra microservices, and network hops
  • ✓ Full ACID compliance, transactional integrity, relational SQL joins, and row-level security on vector tables
  • ✓ Supports exact flat search, approximate IVF-flat, and HNSW graph-based similarity indexes

Limitations & Trade-Offs

  • ✕ HNSW index build times and memory usage on Postgres can be slower than specialized standalone C++/Rust engines
  • ✕ Lacks out-of-the-box multimodal vectorization wrappers or automatic reranking pipelines found in dedicated vector DBs

Supabase Vector Evaluation

Key Advantages

  • ✓ Combines vector search with full Postgres ecosystem: Row Level Security, ACID transactions, and relational joins
  • ✓ Instant auto-generated client APIs (REST, GraphQL, real-time WebSockets) with zero backend boilerplate
  • ✓ Permissive open-source stack with official Docker compose deployment for self-hosting on any VPS

Limitations & Trade-Offs

  • ✕ Free tier projects pause automatically after 7 days of inactivity (resumed upon dashboard login)
  • ✕ Dedicated standalone C++ vector stores can achieve higher raw QPS for multi-million vector datasets

Source Quotes

Direct excerpts from official documentation backing our extracted specifications

pgvector Documentation Proof

PRICING:

"Open source extension under the PostgreSQL License with zero license fees."

FREE_TIER:

"Completely free on any self-hosted or cloud-managed PostgreSQL instance."

SELF_HOSTED:

"Standard extension installable via apt, yum, or Dockerfile in any Postgres environment."

Supabase Vector Documentation Proof

PRICING:

"Pro plan is $25/month including 8GB disk space, 100k monthly active users, and no project pausing."

FREE_TIER:

"Free tier includes 2 free projects, 500MB database space, and up to 50,000 monthly active users."

SELF_HOSTED:

"Supabase is open source and self-hostable with Docker."

Frequently Asked Questions: pgvector vs Supabase Vector

Architectural trade-offs and decision criteria

Is pgvector or Supabase Vector cheaper for developers?

pgvector has a starting price of $0/mo (free), while Supabase Vector starts at $25/mo (hybrid). pgvector includes a free tier (100% Free (PostgreSQL license)). Supabase Vector offers a free tier (500MB database storage (2 free projects)).

Can I self-host pgvector and Supabase Vector in my own infrastructure?

pgvector is self-hostable under the open source license. In comparison, Supabase Vector is available for self-hosting with Apache-2.0 licensing.

Do pgvector and Supabase Vector support the Model Context Protocol (MCP)?

pgvector has native support for MCP servers. Supabase Vector fully supports Model Context Protocol integrations.

Which programming languages have official SDKs for pgvector vs Supabase Vector?

pgvector provides official SDKs in: sql, python, typescript, rust, go. Meanwhile, Supabase Vector supports: typescript, python, dart, swift, go.