PromptBlocks is your central hub for managing AI conversations and prompts. It provides a searchable prompt library, dynamic compositions for structured workflows, and a powerful chat interface with built-in Insights features to track insights from your AI conversations.

Simply sign up, add your OpenRouter API key in your profile settings, and start building your prompt library. You can begin chatting immediately, create parameterized prompts with variables, and build compositions for more structured workflows.

Compositions are reusable sequences of prompts that can include metadata placeholders (like date and time), references to other prompts, and variables that get filled in at runtime. They're perfect for creating structured workflows, daily check-ins, or any repeated AI interactions.

Insights is a feature that tracks key insights from your AI conversations. It allows you to highlight important points, maintain context across chat sessions, and easily reference past discussions and decisions.

You can create prompts with variables using double curly braces (e.g., {{variable_name}}). When you use these prompts, you'll be prompted to fill in the variables, making it easy to reuse prompt structures while customizing the content.

While chatting, type '@' to access your saved prompts and compositions. This lets you quickly inject your carefully crafted prompts into conversations without leaving the chat interface.

Dynamic prompts automatically inject real-time information into your conversations. Using special META: tags, you can include current date, time of day, day of the week, calendar events, todo lists, weather, and more alongside your static prompts. This means your AI interactions are always contextually aware - perfect for daily planning, scheduling assistance, or any workflow that benefits from real-time data.

PromptBlocks works with OpenRouter, allowing you to access various AI models and switch between them mid-conversation. Simply add your OpenRouter API key in your profile settings (stored securely in your browser's local storage, never transmitted to our servers) and you can start using different models to balance speed, cost, and capabilities for each task.