PostgresML
PostgresML is a San Francisco-based startup that builds an open-source machine learning platform embedded directly inside PostgreSQL databases.
Profile
PostgresML provides open-source PostgreSQL extensions that allow users to train and run machine learning models directly inside the database using SQL commands.
PostgresML is a San Francisco-based startup that builds an open-source machine learning platform embedded directly inside PostgreSQL databases. Founded in 2022, the company raised a $4.7 million early-stage venture round in May 2023 from a syndicate of nine investors including Amplify Partners and BoxGroup. As of mid-2026, PitchBook lists the company as having just two employees, making it an unusually lean operation for a database infrastructure project.
The core product is a set of PostgreSQL extensions that let developers train, deploy, and manage machine learning models using standard SQL queries, eliminating the need to move data between a database and a separate ML framework. The GitHub repository, which is the public face of the project, has accumulated 6,800 stars and 362 forks as of mid-2025, with 1,814 total commits. The most recent commit in the dossier dates to July 1, 2025, suggesting development activity has slowed or moved to a private fork.
The company's website and social media presence (LinkedIn, Twitter) are listed but no recent blog posts or product announcements appear in the dossier. No revenue figures, customer names, or headcount beyond the two employees are disclosed in any of the sourced materials. The company operates in the crowded 'ML in the database' space, competing with projects like MindsDB and pgvector, as well as cloud database vendors adding ML capabilities. The lack of any disclosed funding or news activity after May 2023 raises questions about the company's current operational status and growth trajectory.
Who buys this
- PostgreSQL users who want to add ML capabilities without moving data to a separate platform
- Small-to-medium engineering teams looking for a simple, SQL-based ML workflow
- Developers building AI features into applications that already use PostgreSQL
- Data teams in startups that prefer open-source, self-hosted infrastructure
Strengths and what to watch
Strengths
- Tight integration with PostgreSQL means no data movement between database and ML engine, reducing latency and complexity for users already on Postgres.
- Open-source codebase with 6,800 GitHub stars and an active community fork/contribution history, providing a base of grassroots adoption.
- Backed by reputable venture investors including Amplify Partners and BoxGroup, giving the company some financial runway despite a small team.
Watch for
- Extremely small headcount (2 employees per PitchBook) raises questions about the company's ability to support enterprise customers, maintain the codebase, or respond to security issues.
- No disclosed funding or product announcements since May 2023, and the last public GitHub commit was July 1, 2025, suggesting the project may be in maintenance mode or the team has pivoted to a private offering.
- Competitive pressure from better-funded alternatives like MindsDB (which raised $50M+), pgvector (standardized in PostgreSQL), and cloud vendors (AWS SageMaker, Google BigQuery ML) that offer similar in-database ML capabilities.
Key Information
- Industry
- AI Frameworks, Tools & Libraries
- Founded
- 2022
- Headquarters
- San Francisco, USA
Frequently Asked Questions
What is PostgresML and what does it do?
PostgresML is an open-source platform that provides PostgreSQL extensions for training and running machine learning models directly inside the database using SQL commands. It eliminates the need to move data between a database and a separate ML framework.
How does PostgresML let you train machine learning models in PostgreSQL?
PostgresML offers PostgreSQL extensions that allow developers to train, deploy, and manage ML models using standard SQL queries. This tight integration reduces latency and complexity by keeping data within the database, enabling a simpler workflow for PostgreSQL users.
Who is PostgresML designed for?
PostgresML targets PostgreSQL users who want to add ML capabilities without moving data, small-to-medium engineering teams seeking a SQL-based ML workflow, developers building AI features into Postgres applications, and startups preferring open-source, self-hosted infrastructure.
How much funding has PostgresML raised and who are the investors?
PostgresML raised a $4.7 million early-stage venture round in May 2023 from nine investors including Amplify Partners and BoxGroup. The company was founded in 2022 and is headquartered in San Francisco, with no disclosed funding after that round.
How many employees does PostgresML have and what is its GitHub activity?
As of mid-2026, PitchBook lists PostgresML with just two employees. The GitHub repository has 6,800 stars, 362 forks, and 1,814 commits, with the last commit on July 1, 2025, suggesting development may have slowed.
How does PostgresML compare to alternatives like MindsDB and pgvector?
PostgresML competes with MindsDB, which raised over $50 million, and pgvector, now standardized in PostgreSQL. Cloud vendors like AWS SageMaker and Google BigQuery ML also offer in-database ML. PostgresML's small team and lack of recent funding raise concerns about its competitive position.
Sources
- github.com — GitHub repository with 6,800 stars, 362 forks, 1,814 commits, last commit July 1, 2025
- pitchbook.com — Company founded 2022, 2 employees, $4.7M early-stage VC round in May 2023, 9 investors including Amplify Partners and BoxGroup, headquartered in San Francisco
- tracxn.com — Confirms founding year, funding amount, and investor list
- techcrunch.com — Context on the broader 'database-centric ML' startup landscape; note this article is about a different company (DBOS) and is older than the cutoff, included only for market context