January 2025 Summaries

The Neon team has released a Slack app to help users monitor their consumption metrics more easily. The app provides automatic notifications when plan limits are reached for compute hours, storage, or data transfer. Users can install the app and view their usage metrics at any time by running specific commands in a Slack DM. Additionally, users can provide feedback on what other notifications they would like to see in Slack by running `/neon feedback` in a DM.

We encountered a timezone issue when building a parking booking app, where users in Italy were booking slots for the wrong time due to automatic JavaScript timezone conversions. To fix this, we created a plain date-time string without any timezone information and sent it directly to the API and stored it in the database. This approach ensured that the database received the exact date and time input by the user without unintended timezone conversions or metadata. The issue was caused by JavaScript automatically adding timezone metadata to the date, resulting in storing a different date and time in the database than what users selected. We fixed this by sending a plain string to the API and storing it directly in the database.

Navigating timezone issues in JavaScript can be challenging, as illustrated by a problem encountered during the development of a parking booking app where Italian users' bookings were incorrectly stored due to timezone metadata automatically added by JavaScript's Date object. The core issue was that the app stored dates with unintended timezone conversions, resulting in a mismatch between the user's selected time and the stored time in the Neon-hosted database in the US. To resolve this, a plain date-time string without timezone information was created and sent to the API, ensuring that the exact time selected by users was stored accurately. The blog post, part of Neon's Community series, highlights how Neon’s seamless integration, scalability, and free plan for personal projects made it an ideal choice for the app's database needs, allowing the developers to focus on resolving the issue efficiently.

A new Slack app has been introduced to help users monitor their Neon consumption metrics, providing automatic notifications when they reach their plan limits for compute hours, storage, or data transfer. Users can easily install the app by following provided instructions, after which they will receive alerts directly in Slack and can check their usage at any time with a simple command. Although this version of the Slack app is now deprecated as of November 2025, users are encouraged to provide feedback on desired notifications. The announcement, posted by Russ Dias, also invites users to explore other related articles and updates from the Neon community.

RAG is a method that takes a user's question, searches for relevant information, and then combines the retrieved info with the question to prompt an AI chat model. It has two main stages: preparing and indexing information, and dealing with questions. Postgres plays a key role in RAG, often used as a vector database, and pgrag is an extension that helps create a complete RAG pipeline without leaving psql. pgrag provides features such as text extraction, chunking, and embedding generation for both stages of the pipeline. The pipeline uses a vector database to store document chunks and their corresponding embeddings, allowing for efficient searching and retrieval of relevant information. With pgrag, users can create a complete RAG pipeline quickly and easily, using SQL queries instead of writing custom code. The extension is available on Neon right out of the box or can be compiled for use with other Postgres servers.

Neon's open-source extension, pgrag, facilitates building an end-to-end Retrieval-Augmented Generation (RAG) pipeline entirely within PostgreSQL (psql) by integrating features like PDF-to-text conversion, local embedding generation, and AI API requests. The RAG pipeline operates by taking user questions, retrieving relevant information, and using this data to prompt an AI chat model for an informed response. pgrag simplifies the creation of this pipeline by allowing the entire process to be managed with SQL commands, eliminating the need for multiple programming languages and libraries. The extension includes features for text extraction from various file types, splitting text into chunks, generating embeddings using models that can run locally or through third-party APIs, and reranking document chunks to enhance response accuracy. This streamlined approach, which can be deployed directly on the Neon platform or on other PostgreSQL servers, aims to provide an efficient and user-friendly method for setting up RAG pipelines.

AWS is launching a free workshop on using Neon as a development environment for AWS RDS, which will help teams reduce operational overhead and boost productivity. Neon provides an isolated environment for development and testing, allowing teams to automate environment creation and synchronization with production. The workshop will cover topics such as introducing Neon to RDS, setting up a synchronized copy of the database in Neon, automating synchronization workflows, and best practices for efficient workflows. The live event is scheduled for February 27th and can be registered for here.

AWS workshops provide free resources to tackle real-world challenges using AWS infrastructure, and a new workshop focuses on integrating Neon as a development environment with AWS RDS to enhance Postgres development workflows. Neon enables the creation of isolated, ephemeral environments quickly, which reduces operational overhead and allows for seamless integration with AWS RDS without the need for complex production migrations. It provides the ability to spin up and discard environments as needed, ensuring data consistency and enabling automated environment creation via CI/CD pipelines. The workshop covers setting up a synchronized copy of an RDS database in Neon, creating multiple isolated environments using database branching, and automating synchronization processes using GitHub Actions. Additionally, the workshop provides best practices for efficient workflows, emphasizing cost-effectiveness and responsiveness through features like Neon’s scale-to-zero capability.

Instagres is a tool that allows users to spin up a fully functional Postgres database in their browser in under a second, providing instant access and flexibility to transfer the database to a free Neon account when needed. It offers two ways to launch Postgres databases: through the browser or using the CLI with NodeJS. Instagres works by interacting with Neon's API to create a new Postgres database, which can be claimed for long-term use with one click. The tool is built as a Cloudflare Workers app and uses React Router v7 and Drizzle ORM to manage queries. It also includes bot protection measures such as Turnstile CAPTCHA and rate limiting APIs to prevent abuse.

Instagres is a tool that enables users to quickly launch a fully functional Postgres database directly in their browser or through a command line interface, without the need for signing up or setting up infrastructure. Developed using Cloudflare Workers and Drizzle ORM, Instagres provides instant access to a Postgres database by leveraging Neon's serverless platform, which allows for the creation of databases in under a second. Users can initially use these databases temporarily, but the tool offers the option to transfer them to a permanent free account on Neon, ensuring data persistence. Instagres employs Cloudflare’s bot protection and rate limiting features to prevent misuse while maintaining a seamless user experience. This combination of instant database creation and easy transfer to a persistent state makes it useful for prototyping, tutorials, and quick experiments.

The Generative Feedback Loop (GFL) is an approach to optimize and improve the outputs of Large Language Models (LLMs) like GPT. It involves creating a dynamic cycle that adapts LLMs to new and continuously changing data, and user needs. This can be achieved through personalized recommendations, targeted ads, or identifying trends, where an AI might suggest products to a user based on their browsing history, clicks, or purchases. The feedback loop provides benefits such as being dynamic, personalized, and scalable, allowing the model to improve its output without manual intervention. In this article, a sample solution is presented that suggests podcasts to users based on their listening history, using GFL to create a feedback loop that learns user profiles and finds out what they love to listen to. The project uses Azure Serverless Functions in Python to handle real-time requests, stores data in the Neon database, retrieves user listening history, generates vector embeddings for available podcasts, and provides personalized podcast suggestions.

At Neon, the company takes pride in being GDPR-compliant. The General Data Protection Regulation (GDPR) is a comprehensive data protection law that regulates how personal data of EU citizens is collected, processed, and stored. To ensure compliance, organizations must demonstrate accountability through measures such as conducting data protection impact assessments, maintaining detailed records, and appointing Data Protection Officers in certain cases. Neon ensures its platform aligns with GDPR standards by implementing robust security measures, including encryption, access controls, and monitoring and auditing. The company also provides a comprehensive Data Processing Agreement that covers aspects such as data protection safeguards, support for GDPR rights, breach notification protocols, subprocessor accountability, and alignment with GDPR requirements. By choosing Neon, organizations can simplify their path to GDPR compliance, reducing the burden on their own compliance efforts.

Neon emphasizes its commitment to GDPR compliance by offering a platform that adheres to the European Union's data protection regulations, which require organizations to manage and protect personal data effectively. GDPR, introduced in 2018, gives individuals greater control over their personal data and holds companies accountable for data protection, with severe penalties for non-compliance. Neon functions as a subprocessor when customers store personal data on its platform, ensuring compliance through a Data Processing Agreement that includes data protection safeguards, support for GDPR rights, breach notification protocols, and transparency about subprocessors. The company implements security measures such as encryption, access controls, and monitoring to align with GDPR standards, and is ISO 27701 certified, providing customers with a reliable foundation for handling personal data. By choosing Neon, organizations can simplify their path to GDPR compliance, reducing the compliance burden while maintaining robust data security and privacy.

Mocks for databases are often fragile and difficult to maintain, requiring precise control over complex database behavior. They can lead to brittle tests that are tightly coupled to specific mock implementations, making them prone to maintenance issues and debugging challenges. In contrast, branching provides a more reliable and efficient approach by creating isolated copies of the actual database, allowing developers to test with real data and behavior while minimizing overhead. This approach enables faster development cycles, reduced resource costs, and improved reliability, ultimately leading to better system performance and user experience.

Database mocking, particularly for systems like Postgres, is often criticized for being brittle, complex, and providing little confidence in application correctness due to its detachment from actual database behavior. The practice involves replacing real database calls with predictable, fake ones to test application logic, but it often leads to tightly coupled test logic and significant maintenance challenges, especially as the database evolves. Instead, the article suggests using branching as an alternative, which involves creating lightweight, isolated copies of the actual database through copy-on-write technology. This approach offers the benefits of real data and behavior without the overhead of full database copies, allowing for isolated testing environments, realistic test data, and efficient continuous integration processes. Branching is presented as a more reliable and cost-effective solution, providing the isolation and reproducibility promised by mocks, but with real database interactions and data, thereby enhancing reliability, speeding up development cycles, and reducing overhead.

This project aims to build an AI-powered email assistant that can draft, edit, and send emails, store them as vector embeddings, and retrieve similar emails using contextual search. The assistant leverages Semantic Kernel for AI operations, Neon for vector storage, and Azure OpenAI API for generative AI, with everything hosted on Azure. It features AI-powered draft generation, direct email sending, and contextual search capabilities. The project is built using React for the frontend, Semantic Kernel Backend for the backend, and NEON Database for storing vector embeddings. It utilizes various technologies such as .NET, PostgreSQL, Microsoft Graph API, and Azure OpenAI API to provide a secure and efficient email management system.

The `compare_schema` API endpoint allows users to easily compare schemas between Neon databases, including the Free Plan. This feature enables automation of database migrations in agentic systems, which are autonomous AI entities that dynamically determine their actions to achieve complex objectives. The API facilitates this workflow by enabling programmatic schema comparisons, allowing developers to automate the generation and application of migration scripts during deployment. The `compare_schema` API can be used to track schema discrepancies between Neon branches, reducing the risk of errors while supporting a smoother deployment process.

Neon has introduced an API endpoint called compare_schema, which enhances its Schema Diff feature by allowing users to track database schema changes programmatically. This API, which can be integrated into agentic systems and deployment pipelines, enables users to compare schemas between Neon databases, facilitating tasks such as automating schema comparisons, generating migration scripts, and validating database updates. Agentic systems, which are autonomous AI entities capable of dynamic decision-making, benefit from this API for managing database migrations, as it allows them to assess and apply necessary schema updates autonomously. The compare_schema API is also advantageous for developers, as it streamlines the automation of database migration processes by comparing current and target schemas, identifying discrepancies, and testing migrations in isolated Neon branches before deploying them to production, thereby reducing errors and ensuring smoother deployments. The API accepts parameters such as project_id, branch_id, db_name, and optional point-in-time parameters, allowing for detailed schema comparisons between branches.

You've built an Intelligent SQL Query Assistant with Neon Serverless Postgres AI features, utilizing .NET Core for backend logic and API development in C#, Azure Functions to create two serverless APIs: SchemaTraining API and QueryAssistant API, and Azure OpenAI SDK to leverage AI models in .NET code for generating embeddings and translating user queries into SQL. The assistant can process natural language queries, generate SQL commands dynamically, execute them, and return results securely. It also provides control over data handling to comply with standards like GDPR or HIPAA. With this foundation, you can expand the capabilities of your SQL Query Assistant with other Neon features such as Database Branching and Multi-Database Support.

When building an application, teams typically need dedicated development and testing environments. Persistent environments, which mimic production as closely as possible, are often set up but come with significant overhead due to maintenance issues such as configuration drift, seed data maintenance, concurrency problems, and growing costs. In contrast, ephemeral environments are temporary, on-demand environments that can be created and torn down automatically, reducing maintenance overhead. However, a common challenge in ephemeral environments is managing databases that need to populate each environment, which can lead to similar issues as persistent environments. To address this, Neon branches are used to deploy Postgres databases in seconds, with data, offering copy-on-write magic, ready instantly, ephemeral by design, and one-click reset features. By integrating Neon branches into ephemeral environments, teams can solve the last major hurdle holding their speed down – databases.

Building and maintaining traditional persistent environments for software development can be resource-intensive and cumbersome, particularly due to challenges such as configuration drift, seed data maintenance, concurrency issues, and high costs. To address these, teams are increasingly adopting ephemeral environments, which are temporary, on-demand setups that can be automatically created and discarded, offering greater flexibility and integration into CI/CD pipelines. However, replicating production-like databases in these environments often presents a bottleneck due to the time-consuming process of loading seed data. Neon, a serverless Postgres platform, offers a solution by allowing the rapid deployment of Postgres databases using lightweight, copy-on-write branches that reference a single source of truth, enabling quick and efficient data consistency across ephemeral environments. This approach reduces maintenance overhead and costs while supporting scalable and agile workflows.

The developer Neon offers a serverless architecture that enables database branching via copy-on-write, allowing for the creation of ephemeral development environments with isolated data. This feature is particularly useful for teams that need to test and merge frequently, as it saves time and money by eliminating the need for separate testing databases. Branching can be used in various scenarios, including local development, previews, and ephemeral dev/test environments, and Neon's implementation distinguishes itself from other database branching features like Planetscale's Data Branching feature. The company is currently working on a new feature that would allow users to branch directly from production, with only a representative subset of data and PII already transformed.

Neon's innovative branching feature, powered by its serverless architecture with decoupled storage and compute, allows for instantaneous creation of database branches using a copy-on-write mechanism. This technology enables users to create agile, ephemeral branches from a main dataset without reloading the entire dataset, as it references existing data pages and writes new ones only for changes. Such branches are isolated, allowing modifications without affecting the parent data, and can be reset or deleted easily, making them ideal for development, testing, and preview environments. Unlike schema-only branching, which clones only the database structure, Neon's approach includes both schema and data, offering a comprehensive solution for managing large datasets efficiently. Users benefit from reduced costs and time savings, as they can quickly spin up environments without maintaining separate testing databases. Neon is also exploring the next frontier of data branching by incorporating features that allow for selective data transformation, particularly for environments requiring anonymized or representative data subsets, enhancing its practical application in data-sensitive contexts.