Everything on AI.
38 articles tagged AI, sorted newest first.
Classifying Into a Deep Category Tree with Jev and Beam Search in Laravel
A 4,000-leaf taxonomy will not fit in one Choice question, and taking the best child at every level gets you confidently wrong answers. Here is the beam search that fixes it, in PHP.
Gate Your RAG Passages with Jev Before They Reach the LLM
Cosine similarity ranked a prompt injection first and the passage that corrected the question seventh. Here's the Laravel gate I put between retrieval and generation.
Stop Asking the Model for a Verdict: Composite Scoring with Jev in Laravel
A single AI quality score is a number nobody can defend. Here's how to split the judgement into atomic Jev Score questions and combine them with weights your own code owns.
Ask Jev Thirteen Questions at Once: Batching and Speculative Fan-Out in Laravel
Thirteen Jev questions in one request cost roughly what one question costs, because the state is only sent once. Here's the Laravel QuestionSet that makes adding a question a one-line change.
A Laravel Validation Rule That Scores Spam and Toxicity with Jev
Honeypots stop bots, not humans dropping promotional links. Here's a Laravel validation rule that sends the submission to Jev, reads the score and the confidence, and fails open when the API is unreachable.
Check Your LLM's Homework: Verifying Laravel AI SDK Output with Jev
A second LLM grading the first one is slow, expensive and wrong in the same places. Here's how I verify Laravel AI SDK answers with Jev instead.
Route Support Tickets with Jev: Confidence-Gated Triage in Laravel
Send a support ticket to Jev once, get a typed department, an urgency probability and a frustration score back, then let confidence decide whether to assign it or hand it to a human.
Let the Model Look Things Up: WebSearch, WebFetch and FileSearch in the Laravel AI SDK
Three lines of config give your agent the open web and your own documents. Here's how to enable them, which providers actually implement them, and what to lock down before it ships.
Ask a Model About a PDF: Document Attachments in the Laravel AI SDK
A customer uploads a contract and asks a question about it. Here is how to answer it with a document attachment instead of a PDF parser or a week of RAG plumbing.
Stop Shipping Prompt Regressions: LLM Evals for the Laravel AI SDK with Pest 5
Your faked-agent tests prove the wiring works and tell you nothing about whether the output is still any good. Pest 5's evals plugin closes that gap.
Generate Speech from Text in Laravel with the AI SDK's Audio Class
The Laravel AI SDK turns text-to-speech into four lines and hands you a Storage path instead of a byte string. Here is the whole surface — voices, queueing, serving — plus the hash cache that stops you paying twice for the same sentence.
Transcribe Audio in Laravel with the AI SDK: Uploads, Storage and Speaker Diarization
Transcription ships in the Laravel AI SDK with queueing, diarization and faking built in. Almost nobody knows it's there, so here's the whole path.
Generate and Store Images with the Laravel AI SDK
The AI SDK docs lead with text, so most people reach for a raw HTTP client to generate images. There's a first-class API for it — here's how I wire it up.
Manage Laravel AI SDK Prompts as Versioned Blade Templates
Heredoc prompts buried in service classes cannot be diffed, reviewed or rolled back. Here is how I move them into versioned Blade templates that CI can catch.
Add Input and Output Guardrails to a Laravel AI SDK Agent
The Laravel AI SDK ships no moderation API, but it does give you one clean interception seam. Here is how I wrap an agent with input and output guardrails using agent middleware.
Cut LLM Costs with a Semantic Response Cache in the Laravel AI SDK
Exact-match caching never fires on natural-language prompts. Here's how I cache LLM responses by meaning using embeddings, pgvector, and the Laravel AI SDK.
Build Multi-Agent Workflows with the Laravel AI SDK
One agent with a dozen tools and a bloated prompt gets flaky fast. Here's how to split the work across a coordinator and focused specialists with the Laravel AI SDK.
Give Laravel AI SDK Agents External Tools with MCP Servers
Stop hand-writing a tool for every capability. Register an MCP client, authenticate it, and spread its tools straight into your Laravel AI SDK agent.
Run Laravel AI SDK Agents in the Background with Queued Jobs
A synchronous AI call can tie up a PHP worker for half a minute. Here's how to run Laravel AI SDK agents on a queue and surface the result when it's done.
Persist Multi-Turn Chat History with the Laravel AI SDK
A stateless agent forgets every turn. Here's how to wire persistent conversation memory into a Laravel AI SDK agent, resume a chat by ID, and stop long histories from wrecking your token budget.
Track Token Usage and Cost with the Laravel AI SDK
Every Laravel AI SDK response carries a usage object. Here's how to capture it, attribute it to a user and feature, and turn tokens into a real cost dashboard.
Laravel AI SDK Tool Calling: Let the Model Run Your PHP
Give a Laravel AI agent real abilities: typed PHP tools the model can call to look up data, with enum-constrained inputs and a hard cap on runaway loops.
Test AI Features Without Calling the API: Agent::fake() in the Laravel AI SDK
You shipped an agent and wrote zero tests for it, because hitting a real model is slow, costs tokens, and needs a key in CI. The Laravel AI SDK's fake() fixes all three.
Analyze Images with the Laravel AI SDK: Vision and Multimodal Prompts
The Laravel AI SDK does more than generate text. Here's how to attach an image to a prompt and have a vision model describe, classify, or extract data from it.
Provider Fallback and Failover with the Laravel AI SDK
A single LLM provider is a single point of failure. Here's how to fail over to a backup provider with the Laravel AI SDK without rewriting a line of prompt code.
Laravel AI SDK Structured Output — JSON Schema Enforcement for LLM Responses
Stop parsing markdown fences and stripping prose out of LLM replies. The Laravel AI SDK ships a HasStructuredOutput contract that forces the model to return JSON matching the schema you describe in PHP — and the same agent works across every supported provider.
Stream Prism AI Responses Token-by-Token into Livewire with wire:stream
A ChatGPT-style chat UI in Laravel without writing a line of JavaScript, without WebSockets, without a separate Node process. Prism streams the tokens, Livewire's wire:stream paints them, you write the bubble.
Trace Every LLM Call: Langfuse Observability for Prism in Laravel
Prism gives you the LLM API; Langfuse gives you the receipts. Wire them together with one env var and trace every prompt, token, and tool call in twenty minutes.
Production AI Agents in Laravel: Reliable Loops with Prism and Laravel Workflow
A production-grade playbook for Laravel AI agents: verifier loops, idempotent tools, durable workflow state, Langfuse tracing, and evaluating agent quality in CI.
Laravel 13 whereVectorSimilarTo(): Native Semantic Search in the Query Builder
Laravel 13's whereVectorSimilarTo() turns the query builder into a vector store. Plain strings in, ranked results out: no manual embedding pipeline.
Building a RAG Pipeline with Laravel AI SDK and pgvector
Build a complete retrieval-augmented generation pipeline in Laravel: from embedding documents and storing vectors in pgvector to generating grounded answers with the AI SDK.
Laravel Boost MCP: Let AI Agents Understand Your Codebase
AI coding assistants stumble on Laravel's magic methods and facades. Boost is a local MCP server that gives them structured context so they stop guessing.
Laravel MCP Server: Expose Your App to AI Clients in 15 Minutes
The laravel/mcp package turns your Laravel app into an MCP server with a few artisan commands: no Node, no Python, just familiar middleware and routing.
Laravel Prism Tool Calling: Build AI Agents That Actually Do Things
Prism's tool calling turns a text-generating LLM into an agent that can actually *do* things: query your database, hit an external API, or dispatch a queued job.
Upgrading From Laravel 12 to Laravel 13: The Practical Guide
The Laravel 13 upgrade is mostly painless: but three config areas will silently change how your app behaves if you skip them. Here's what actually changed and what's worth adopting.
The Complete Guide to the Laravel AI SDK in Laravel 13
Laravel 13 ships a first-party AI SDK that replaces the patchwork of vendor SDKs and community packages. Here's how to use every capability: agents, streaming, structured output, tools, embeddings and RAG: in one working project.
Getting Started With Laravel Prism: Add AI to Any Laravel App in Minutes
Prism gives you a provider-agnostic fluent interface for OpenAI, Anthropic, Ollama, and more inside any Laravel app. Here's how to go from zero to your first AI response.
Building an Automated Bug-Fixing Pipeline with Claude Cowork
A Claude Cowork scheduled task that reads Sentry errors, triages them via Linear, implements fixes in your Laravel codebase, and opens a GitHub PR: without you lifting a finger until review.