PHPackages noah-medra/prompt-builder - PHPackages - PHPackages [Skip to content](#main-content)[PHPackages](/)[Directory](/)[Categories](/categories)[Trending](/trending)[Leaderboard](/leaderboard)[Changelog](/changelog)[Analyze](/analyze)[Collections](/collections)[Log in](/login)[Sign up](/register) 1. [Directory](/) 2. / 3. [Utility & Helpers](/categories/utility) 4. / 5. noah-medra/prompt-builder ActiveLibrary[Utility & Helpers](/categories/utility) noah-medra/prompt-builder ========================= Laravel package to build dynamic AI prompts 1.0.2(5mo ago)037↓100%MITPHPPHP ^8.0 Since Dec 2Pushed 5mo agoCompare [ Source](https://github.com/oscarmedra/prompt-builder)[ Packagist](https://packagist.org/packages/noah-medra/prompt-builder)[ RSS](/packages/noah-medra-prompt-builder/feed)WikiDiscussions main Synced 1mo ago READMEChangelogDependencies (7)Versions (4)Used By (0) PromptBuilder ============= [](#promptbuilder) **PromptBuilder** is a Laravel package that allows you to create and execute AI prompts in a flexible and structured way. This package enables you to generate dynamic queries, customize parameters, and manage conversation history to improve the relevance of responses. Installation ------------ [](#installation) ### Composer [](#composer) To install **PromptBuilder** via Composer, run the following command: ``` composer require noahmedra/promptbuilder ``` Usage ----- [](#usage) ### Creating a `PromptBuilder` object [](#creating-a-promptbuilder-object) Start by instantiating a `PromptBuilder` object: ``` use NoahMedra\PromptBuilder\PromptBuilder; $builder = PromptBuilder::make(); ``` ### Defining the **Driver** (Processing Engine) [](#defining-the-driver-processing-engine) By default, **PromptBuilder** uses the `OllamaDriver`. If you want to use another driver, you can specify it using the `driver()` method. ``` use NoahMedra\PromptBuilder\Drivers\HuggingFaceDriver; $builder->driver(HuggingFaceDriver::class); ``` ### Adding **instructions** (with Sub-Instructions and Conditions) [](#adding-instructions-with-sub-instructions-and-conditions) Instructions are rules or constraints you want to apply to the AI's generated response. These instructions can have **sub-instructions** or be **condition-dependent**. You can chain instructions infinitely and even create **nested conditions**. Here is an example where instructions can contain sub-instructions, and conditional logic can be applied using the `when()` method. #### Example with Instructions and Sub-Instructions: [](#example-with-instructions-and-sub-instructions) ``` $builder->instruction("### Financial History") ->instruction("We will also review your financial history to identify trends.") ->when( true, // This condition is true, so the sub-instructions will be applied function($builder) { $builder->instruction("Here is a summary of your financial history for the last three months."); ->when(false, fn($ist) => $ist->add('It seems there are no significant changes in your financial behavior this month. We’ll continue to monitor this trend for future insights.')) $builder->instruction("Your balance has fluctuated between X and Y. It seems like you had some unexpected expenses."); } ); ``` In this example: 1. **Main instructions**: The first two instructions introduce the financial history topic and explain the review process. 2. **Conditional instructions**: The `when()` method checks a condition (in this case, `true`), and if true, additional **sub-instructions** are added (e.g., a summary of the financial history). ### Adding **dynamic parameters** [](#adding-dynamic-parameters) You can pass custom parameters to your prompt with the `withParams()` method: ``` $builder->withParams(['key' => 'value']); ``` ### Defining a **context** [](#defining-a-context) You can add additional context to guide the AI in generating the response. ``` $builder->context("You are a virtual assistant with expertise in financial analysis."); ``` ### Asking a **question** [](#asking-a-question) You can ask a question that will be included in the final prompt. ``` $builder->ask("What are the main trends in the past three months of my financial data?"); ``` ### Managing **history** [](#managing-history) If you want the AI to use previous conversation history to provide more contextual responses, enable the `useHistory()` option. ``` $builder->useHistory(true); ``` ### Generating and retrieving the **response** [](#generating-and-retrieving-the-response) Once all instructions and parameters are defined, you can generate the prompt and obtain the AI's response by calling the `getOutput()` method: ``` $builder->process(); $output = $builder->getOutput(); echo $output->('message.content'); // Display the generated response echo $output->('model'); // Display the model's response ``` ### Handling **JSON** responses [](#handling-json-responses) If you want the AI's response to be formatted as JSON, use the `expectResponseFormat()` method. ``` $builder->expectResponseFormat('{"resume": "Summary of the response", "response": "Your response here"}'); ``` ### **Driver: How It Works (Input/Output)** [](#driver-how-it-works-inputoutput) The `Driver` is responsible for handling the input (your prompt) and generating the output (the AI's response). Each driver implements the `DriverInterface` and provides a `process()` method that takes a `BuilderInput` object, processes the prompt, and returns a `BuilderOutput`. For example, the `OllamaDriver` sends the prompt to a local Ollama service and receives the AI's response. Here is an example of the **OllamaDriver**: ``` namespace App\Drivers; use Exception; use Illuminate\Support\Facades\Http; use NoahMedra\PromptBuilder\BuilderInput; use NoahMedra\PromptBuilder\BuilderOutput; use NoahMedra\PromptBuilder\Drivers\DriverInterface; class OllamaDriver implements DriverInterface { public function process(BuilderInput $input) : BuilderOutput { return new BuilderOutput($this->executePrompt($input)); } private function executePrompt(BuilderInput $input) : string { $response = Http::withHeaders([ 'Content-Type' => 'application/x-www-form-urlencoded', ])->post('http://localhost:11434/api/chat', [ 'model' => 'llama3.1', 'stream' => false, 'messages' => [ [ 'role' => 'user', 'content' => $input->getPromptText(), ] ], ...($input->getParams() ?? []) ]); if ($response->failed()) { throw new Exception($response->body()); } return $response->body(); } } ``` In this example: 1. The `process()` method receives a `BuilderInput` object which contains the prompt text. 2. The `executePrompt()` method sends the prompt to a remote API (Ollama) and receives the AI response. 3. The response is returned as a `BuilderOutput` object. ### Complete Example [](#complete-example) Here is a complete example of how to use **PromptBuilder**: ``` use NoahMedra\PromptBuilder\PromptBuilder; $builder = PromptBuilder::make(); // Add parameters, instructions, context, and a question $builder->withParams(['lang' => 'en']) ->instruction("Provide a clear and concise answer.") ->instruction("The response should be in JSON format.") ->instruction("Include the following sub-instructions:") ->when(true, function($builder){ $builder ->instruction("Ensure the answer is clear and easy to understand.") ->instruction("Be concise and avoid unnecessary details.") }) ->context("You are an expert in financial analysis.") ->ask("What are the main trends in the past three months of my financial data?"); // Generate and get the response $builder->process(); $output = $builder->getOutput(); echo $output->get('message.content'); ``` Features -------- [](#features) - **Driver flexibility**: Easily add or replace text-processing engines. - **Dynamic instructions with sub-instructions and conditions**: Add custom instructions that can have nested sub-instructions, and create complex prompts with **conditional instructions** using the `when()` method. - **History management**: Store previous conversations to provide context. - **JSON response support**: Format the AI's responses in JSON for easy integration. - **Dynamic parameters**: Pass custom parameters to your prompts. Customization ------------- [](#customization) You can easily extend **PromptBuilder** by creating your own **drivers** or by customizing the **PromptDriverInterface**. To do this, simply create a class that implements the `PromptDriverInterface` and pass it to the `driver()` method. Contributions ------------- [](#contributions) Contributions are welcome! If you have ideas to improve the package, feel free to submit an issue or a pull request. ### Health Score 34 — LowBetter than 77% of packages Maintenance72 Regular maintenance activity Popularity8 Limited adoption so far Community6 Small or concentrated contributor base Maturity42 Maturing project, gaining track record Bus Factor1 Top contributor holds 100% of commits — single point of failure How is this calculated?**Maintenance (25%)** — Last commit recency, latest release date, and issue-to-star ratio. Uses a 2-year decay window. **Popularity (30%)** — Total and monthly downloads, GitHub stars, and forks. Logarithmic scaling prevents top-heavy scores. **Community (15%)** — Contributors, dependents, forks, watchers, and maintainers. Measures real ecosystem engagement. **Maturity (30%)** — Project age, version count, PHP version support, and release stability. ### Release Activity Cadence Every ~0 days Total 3 Last Release 161d ago ### Community Maintainers ![](https://www.gravatar.com/avatar/1e803756b25f301e910c690b555b74f409327d97e4162284ed90a332a45d3b80?d=identicon)[oscarmedra](/maintainers/oscarmedra) --- Top Contributors [![oscarmedra](https://avatars.githubusercontent.com/u/53277746?v=4)](https://github.com/oscarmedra "oscarmedra (47 commits)") ### Code Quality TestsPest Code StyleLaravel Pint ### Embed Badge ![Health badge](/badges/noah-medra-prompt-builder/health.svg) ``` [![Health](https://phpackages.com/badges/noah-medra-prompt-builder/health.svg)](https://phpackages.com/packages/noah-medra-prompt-builder) ``` ### Alternatives [stevebauman/location Retrieve a user's location by their IP Address 1.3k7.6M65](/packages/stevebauman-location)[nativephp/mobile NativePHP for Mobile 82724.0k43](/packages/nativephp-mobile)[bensampo/laravel-embed Painless responsive embeds for videos, slideshows and more. 142146.8k](/packages/bensampo-laravel-embed)[glhd/conveyor-belt 14797.0k](/packages/glhd-conveyor-belt)[pulkitjalan/ip-geolocation IP Geolocation Wrapper with Laravel Support 89164.9k1](/packages/pulkitjalan-ip-geolocation)[flarum/core Delightfully simple forum software. 211.3M1.9k](/packages/flarum-core) PHPackages © 2026 [Directory](/)[Categories](/categories)[Trending](/trending)[Changelog](/changelog)[Analyze](/analyze)