LLMs are great at thinking, but terrible at doing. Learn how Tool Calling bridges the gap between text generation and real-world action.

Tool Calling in LLMs: How Models Talk to the Real World

Announcing NodeLLM, the Backend-First infrastructure layer for building production-grade AI agents in Node.js. Provider-agnostic, zero-dependency, and built for long-running processes.

Introducing NodeLLM: The Backend-First AI SDK for Node.js

Efficiently preload has_many associations in GraphQL Ruby to eliminate N+1 queries using the graphql-preload gem.

preload has_many associations in graphql-ruby using graphql-preload

Improve your GraphQL mutations by using InputObjectType to group and clean up mutation arguments.

GraphQL Ruby Mutation with input object type

Recap of my first experience speaking at a major conference, Deccan RubyConf 2017, held in Pune.

Deccan RubyConf 2017 - First Experience as a speaker

NodeLLM introduces native support for the Model Context Protocol (MCP), providing a standardized boundary for discovering and executing tools, resources, and prompt templates across any compliant server.

Standardizing Agent Connectivity with Model Context Protocol (MCP)

NodeLLM 1.15 introduces automated schema self-correction, enabling agents to recover from validation failures without manual intervention. This release also brings fine-grained middleware lifecycle control and hardened type safety.

NodeLLM 1.15: Automated Schema Self-Correction and Middleware Lifecycle

Track LLM costs, latency, and errors with a self-hosted dashboard. No SaaS fees, no data leaving your infrastructure.

NodeLLM Monitor: Self-Hosted LLM Observability

NodeLLM 1.14 reinforces our philosophy that agents don't require complex orchestration frameworks—they are just LLMs armed with tools. This release also brings first-class support for xAI (Grok) and the complete Mistral suite.

NodeLLM 1.14: Demystifying Agents and Expanding the Ecosystem

Vercel AI SDK is the industry standard for shipping fast. NodeLLM is a backend-first LLM runtime with middleware architecture, 540+ models across 7 providers, and enterprise-grade features.

Vercel AI SDK and NodeLLM: Choosing the Right Layer

NodeLLM 1.10 introduces a full middleware architecture for intercepting LLM requests, responses, tool executions, and errors. Build PII protection, cost guards, and custom pipelines—without changing your business logic.

NodeLLM 1.10: Production-Grade Middleware System

Introducing @node-llm/monitor—a production-grade observability layer for LLM applications. Track costs, latency, token usage, and debug AI interactions with a built-in real-time dashboard.

NodeLLM Monitor: Production Observability for LLM Applications

Testing AI systems is often frustrating and expensive. I’ve been working on a small utility to make testing LLM interactions a bit more predictable and secure—here is `@node-llm/testing`.

Building a more predictable way to test LLMs: Introducing @node-llm/testing

Prompt injection is the new SQL injection. This post explores the security architecture required for 2026 AI Agents and how NodeLLM provides the necessary guardrails.

The Invisible Perimeter: Hardening LLM Flows in the Age of Autonomous Exploits

Infrastructure doesn't stop at OpenAI. Learn how to extend NodeLLM to support proprietary gateways like Oracle Cloud's Generative AI Service using a clean, zero-dependency, interface-driven approach.

Beyond the Big Three: Building Custom NodeLLM Providers for Enterprise Cloud Gateways

NodeLLM 1.7.0 introduces a unified interface for reasoning-focused models. Standardize how you handle Claude 3.7 Thinking, DeepSeek R1 Reasoning, and OpenAI o1/o3 Effort with a single, fluent API.

NodeLLM 1.7.0: Standardizing 'Extended Thinking' Across Providers

AI Agents are more than chatbots—they're systems that can reason, plan, and act. This post breaks down what an agent actually is, how the execution loop works under the hood, and gives you a copy-paste-ready agent using NodeLLM.

Building Your First AI Agent in Node.js: A Deep Dive

Every time I start a new AI project, I find myself redesigning the same four database tables. Here is why we need a universal persistence layer for AI, and how @node-llm/orm brings Rails-like sanity to the Node.js AI ecosystem.

Stop Redesigning Chat Databases: The Case for a Universal AI ORM

Building a chatbot that lists models is easy. Building one that remembers users, persists tool calls, and handles real-time streaming in a production Next.js app is hard. Here is how we do it with @node-llm/orm.

Stop Building Toy Chatbots: A Guide to Streaming, Persistence, and RAG in Node.js

NodeLLM 1.5.0 adds runtime safety for AI integrations: mandatory timeouts, tool call limits, human-in-the-loop approvals, and global token protection.

NodeLLM 1.5.0: Putting Security in the Driver's Seat

Protect your GraphQL server from DDoS attacks and resource exhaustion by implementing timeouts, max depth, and complexity limits.

GraphQL Security Issues and Solutions

Use GraphQL Interfaces to share common fields across different types and make your schema definitions more maintainable.

DRY GraphQL definitions using interfaces

Learn how to effectively handle and expose errors in your GraphQL Ruby API, from validation failures to unexpected exceptions.

GraphQL ruby error handling

Mutation is a special type of query used to change data in the database like Creating, Editing or Deleting Records from a table or Store. These are th...

GraphQL -Mutation query implementation - Ruby on Rails

Explore powerful alternatives to the default GraphiQL IDE for better query management and header support.

GraphiQL IDE Alternatives for Development/Testing/Exploring GraphQL Servers

Solve the common N+1 SQL query problem in GraphQL Ruby applications using the graphql-batch gem.

Solving N+1 query in GraphQL using graphql-batch

Learn how to implement nested data sets in GraphQL to fetch complex data structures in a single request.

GraphQL Server - Nested Query Implementation Example in Ruby on Rails

Introduction to implementing a GraphQL server in Ruby on Rails, covering schemas, types, and queries.

A GraphQL server implementation - Ruby On Rails

Monitor your background jobs with a Resque-inspired web interface for delayed_job.

delayed_job_web - A resque inspired interface for delayed_job

Fine-tune your API performance by profiling endpoints with the rack-mini-profiler gem.

Profile API endpoints Using rack-mini-profiler

An introduction to Active Record basics and core CRUD operations in Rails.

slide-Active Record Basics

Learn how to record and store API examples in JSON format using Apipie-rails.

Apipie - Dumping Examples into apipie_examples File

Rails 5 introduces a change where belongs_to associations are required by default. Learn how to manage this new behavior.

Rails 5 'belongs_to' associations default to required true

Camaleon CMS is a powerful open-source alternatives to WordPress, built specifically for Ruby on Rails.

Camaleon CMS - Wordpress Alternative built in Ruby on Rails 4

Setup a new Rails application using the edge version to test new features like Action Cable and API mode.

Create Rails Application from Edge Version -- 5.0.0.beta1.1

Learn how to integrate Travis CI with your GitHub repository for automated builds and testing.

Travis CI integration - Step by step tutorial

A session covering the essentials of Active Record Migrations in Ruby on Rails.

Slide - Playing with Active Record Migrations

Introduction to the basic concepts of Active Record Associations in Ruby on Rails.

Slide - An introduction to Active Record Associations

Introduction to Ruby on Rails slides for beginners and new joiners.

Slide - An Introduction To Ruby On Rails

Speed up your development by running a specific subset of tests using RSpec's tagging feature.

Rspec Run specific set of tests using --tag option

Learn how to implement dependent select boxes in ActiveAdmin to populate options based on previous selections.

Implement Dependent select in Active Admin

Brakeman is a static analysis tool that checks Ruby on Rails applications for security vulnerabilities without needing a full execution stack.

Ruby on Rails applications security vulnerability checking tool - Brakeman

The Ruby I18n gem provides an easy-to-use framework for translating your application's models, attributes, and validation messages.

Translate Active Record Model Name, Attributes, and Error Messages with Rails Internationalization (I18n) API

Solve CORS issues in development by setting up a reverse proxy or disabling browser security checks.

Two Ways to Avoid Cross-Origin Resource Sharing Issues in Development Environment

Active Admin is a Ruby on Rails plugin for generating administration style interfaces. ActiveSkin is a beautiful flat skin for it.

ActiveSkin - Flat skin for active admin

Maintain a consistent order of associations, scopes, and methods in your Rails models for better readability.

Standard for the Order of Associations, Scopes, Includes, and Validations in a Rails Model

SimpleCov is a code coverage analysis tool for Ruby. Set it up with just a few lines of code to get detailed coverage reports.

Check Test Coverage With SimpleCov Gem

Implement banking-grade password policies in Rails using the Devise gem and Devise Security Extension.

Password Policy Implementation with Devise and Devise Security Extension

Document your RESTful API in Rails using Apipie, a DSL that lets you describe your code through the code itself.

Apipie-rails - Rails API documentation tool

Customize Spree Commerce models by adding extra fields using decorators and Deface.

Spree - Add extra fields to product model with deface and decorator

Make your Refinery CMS site multilingual using the refinerycms-i18n gem.

Multiple language sites in Refinery CMS with i18n

HTML5 allows us to implement client-side form validation without coding. Learn how to customize the error messages with Civem.js.

Custom HTML5 Validation Error Messages with Civem.js

OpenShift is a Platform as a Service (PaaS) from RedHat. Learn how to deploy a Rails application to OpenShift gears.

Deploying Ruby On Rails Application To Cloud Application Platform - Openshift

Add in-place editing to ActiveAdmin pages using the Best In Place gem for a seamless user experience.

In Page Editing In 'Active Admin' - Using Gem 'Best In Place'

Implement nested forms for belongs_to relationships in ActiveAdmin using accepts_nested_attributes_for.

Nested forms for belongs_to relationship - ActiveAdmin

FullCalendar is a jQuery plugin that provides a full-sized drag-and-drop calendar. Learn how to integrate it with a Rails application.

FullCalendar - Integration with Ruby On Rails

Setup a has_one :through relationship filter in ActiveAdmin with Project, Task, and Activity models.

ActiveAdmin- adding a filter on has_one through association

Implement autocomplete in the active_admin interface to select one or more users from the user list using the Chosen library.

Autocomplete using Chosen in Rails Active_admin

Reduce Kannel logs by changing the log level or setting up log rotation with logrotate.

Changing Kannel Log Level & Log Rotation

SQL box is a special Kannel box that sits between bearer box and SMS box and uses a database queue to store and forward messages.

Installing and configuring SQL-Based queue engine for Kannel (SQL box)

Over the past few months, I have been facing many issues with the Kannel SMS gateway. I decided to write a tool to monitor Kannel activities.

Kannel monitoring tool - Ruby Gem

Writing a Ruby Gem was a long-time wish of mine. This generator automatically generates custom code for handling 404 and 500 errors and pages in Rails applications.

A Gem for generating Custom error pages in Rails Application

To support MySQL, we need to recompile Kannel with --mysql. To store DLR information in a MySQL database, you can use the dlr-storage = mysql configuration directive.

Configuring MySQL DLR storage

Kannel is an open-source WAP and SMS gateway for GSM networks. Learn how to install and configure it to link HTTP-based services with SMS centres.

All Posts

Tool Calling in LLMs: How Models Talk to the Real World

Introducing NodeLLM: The Backend-First AI SDK for Node.js

preload has_many associations in graphql-ruby using graphql-preload

GraphQL Ruby Mutation with input object type

Deccan RubyConf 2017 - First Experience as a speaker