# Vizra

## Docs

- [Vizra Evals](https://docs.vizra.ai/index.md): Write agent evals as Pest tests. Keep every result — scores, judge reasoning, cost — with baselines, regression gates, and a dashboard.
- [Quickstart](https://docs.vizra.ai/evals/quickstart.md): From install to a baseline-gated, CI-ready agent eval in five minutes.
- [Installation](https://docs.vizra.ai/evals/installation.md): Requirements, package setup, and what gets added to your application.
- [How It Works](https://docs.vizra.ai/evals/how-it-works.md): The lifecycle of an eval run — from dataset rows to recorded, baseline-compared results.
- [Scoring Model](https://docs.vizra.ai/evals/scoring.md): How samples become scores, scores become rates, and rates become a pass or fail.
- [Baselines & Regressions](https://docs.vizra.ai/evals/baselines-and-regressions.md): Every run compared against the run you trusted — row by row, with receipts.
- [Writing Pest Evals](https://docs.vizra.ai/evals/pest.md): The toPassEval() expectation — everything is a Pest test, and the recorder works underneath.
- [Datasets](https://docs.vizra.ai/evals/datasets.md): JSONL, CSV, arrays, Eloquent — or your actual production traffic.
- [Multi-Turn Conversations](https://docs.vizra.ai/evals/multi-turn.md): Evaluate agents in the middle of a conversation — including conversations from production.
- [Assertions Reference](https://docs.vizra.ai/evals/assertions.md): Every deterministic check — from substrings to real tool calls, token cost, and multi-turn behavior.
- [The LLM Judge](https://docs.vizra.ai/evals/judge.md): Structured-output scoring with reasoning you keep — no regex parsing, no black boxes.
- [Judge Calibration](https://docs.vizra.ai/evals/calibration.md): Measure how much your LLM judge agrees with humans — before you let it gate a release.
- [Class-Based Evaluations](https://docs.vizra.ai/evals/class-based-evaluations.md): The full authoring surface — reusable suites, model matrices, and transform hooks.
- [Testing Without Tokens](https://docs.vizra.ai/evals/testing-without-tokens.md): Develop and test your evals end-to-end with zero network calls, using the Laravel AI SDK's own fakes.
- [CLI Commands](https://docs.vizra.ai/evals/cli.md): The same engine without Pest — runs, baselines, scaffolding, and machine-readable output.
- [Running in CI](https://docs.vizra.ai/evals/ci.md): Fail the build when your agent regresses — with cost kept firmly under control.
- [Configuration](https://docs.vizra.ai/evals/configuration.md): Judge defaults, gates, comparison tolerance, concurrency, and the price table behind cost tracking.
- [Dashboard](https://docs.vizra.ai/evals/dashboard.md): Score trends, per-sample drill-downs with judge reasoning, and run comparison — in your browser.
- [Migrating from Vizra ADK](https://docs.vizra.ai/evals/migrate-from-vizra-adk.md): Moving from the ADK's CSV-based evaluations to recorded, baseline-compared evals.
- [Vizra Evals vs pest-plugin-evals](https://docs.vizra.ai/evals/vs-pest-plugin-evals.md): Two expectation sets, one test file — quick checks in one, recorded measurement in the other.
- [Vizra Cloud](https://docs.vizra.ai/cloud/index.md): Your team's eval history in one place — laptop runs, CI runs, and the ability to run a suite from the browser.
- [Connect a Project](https://docs.vizra.ai/cloud/connect.md): Create a project, mint a key, add two lines to .env — and your next run appears.
- [Reporting Runs](https://docs.vizra.ai/cloud/reporting.md): Which surfaces push to Cloud, what's in the payload, and why a flaky uplink can never fail your build.
- [Running from the Dashboard](https://docs.vizra.ai/cloud/triggering.md): A Run button in the browser that executes on your infrastructure — via your scheduler, or your CI.
- [Editing Datasets](https://docs.vizra.ai/cloud/datasets.md): Fix a test question in the browser, run the variant, and commit it back to your repository.
- [Plans & Retention](https://docs.vizra.ai/cloud/plans.md): What each tier includes, and exactly what ages out of your history.
- [Vizra ADK](https://docs.vizra.ai/adk/index.md): Build incredible intelligent agents with the most developer-friendly Laravel package. Your journey to AI mastery starts here!
- [Getting Started](https://docs.vizra.ai/installation/getting-started.md): Launch your first AI agent in minutes with Vizra ADK! Ready to feel the magic? Let's get you building intelligent agents faster than you can say 'artificial intelligence'!
- [Requirements Checklist](https://docs.vizra.ai/installation/requirements.md): Let's make sure your environment is ready for the exciting journey ahead!
- [Configuration](https://docs.vizra.ai/installation/configuration.md): Customize your perfect AI agent setup with Vizra ADK's flexible configuration system
- [Architecture](https://docs.vizra.ai/concepts/architecture.md): Understand the technical architecture of Vizra ADK - how agents, tools, context, and services work together
- [AI Agents](https://docs.vizra.ai/concepts/agents.md): Meet your new AI workforce. Agents are the heart and soul of Vizra ADK - intelligent, conversational, and incredibly powerful.
- [Tools](https://docs.vizra.ai/concepts/tools.md): Supercharge your agents with powerful capabilities. Tools let your AI agents interact with the real world - from databases to APIs to custom business logic.
- [Sessions & Memory](https://docs.vizra.ai/concepts/sessions-memory.md): Give your agents a memory that lasts. Learn how to build agents that remember conversations, learn from interactions, and provide personalized experiences over time.
- [Specialized Agents](https://docs.vizra.ai/agents/overview.md): Discover specialized agent types for audio, images, and more
- [Audio Agent](https://docs.vizra.ai/agents/audio-agent.md): Generate speech and audio with the AudioAgent
- [Image Agent](https://docs.vizra.ai/agents/image-agent.md): Generate images from text descriptions using AI models like DALL-E and Imagen
- [Toolbox](https://docs.vizra.ai/tools/toolbox.md): Group related tools together for better organization and authorization
- [Tool Pipelines](https://docs.vizra.ai/tools/tool-pipelines.md): Chain tools together for powerful sequential execution with data transformation
- [Structured Output](https://docs.vizra.ai/tools/structured-output.md): Get typed, validated responses from your agents with automatic schema enforcement
- [Workflows](https://docs.vizra.ai/concepts/workflows.md): Orchestrate multiple agents together in powerful patterns
- [Vector Memory & RAG](https://docs.vizra.ai/concepts/vector-rag.md): Unlock the power of semantic search and intelligent information retrieval. Let your agents access vast knowledge bases and provide incredibly accurate, context-aware responses.
- [Dynamic Prompts](https://docs.vizra.ai/concepts/dynamic-prompts.md): Master your AI's voice with dynamic prompts. Track improvements over time with versions, test different personalities with variants, and evolve your agents without touching code.
- [Agent Queuing](https://docs.vizra.ai/advanced/agent-queuing.md): Queue agents for background and async processing with Laravel queues
- [Macros & Mixins](https://docs.vizra.ai/advanced/macros.md): Extend Vizra ADK with custom methods using Laravel's powerful macros and mixins pattern
- [LLM Providers](https://docs.vizra.ai/integrations/providers.md): Choose your AI's brain. From lightning-fast Groq to privacy-focused local models, Vizra ADK supports 10 different LLM providers via Prism PHP. Find the perfect match for your use case.
- [OpenAI Compatibility](https://docs.vizra.ai/integrations/openai-compatibility.md): Use your Vizra agents with any OpenAI-compatible tool or library. Drop-in replacement for OpenAI's Chat Completions API that opens up your agents to the entire OpenAI ecosystem.
- [MCP Integration](https://docs.vizra.ai/integrations/mcp.md): Supercharge your agents with Model Context Protocol. Connect to filesystems, databases, APIs, and the entire MCP ecosystem with just one line of code.
- [Laravel Boost Integration](https://docs.vizra.ai/integrations/laravel-boost.md): Use Laravel Boost's AI-powered development assistant to generate Vizra ADK agents with full understanding of the framework's capabilities.
- [Evaluations](https://docs.vizra.ai/testing/evaluations.md): Transform your AI agents from unpredictable to reliable. Test, validate, and perfect your agents with automated quality assurance.
- [Tracing](https://docs.vizra.ai/testing/tracing.md): Get X-ray vision into your agents. See exactly what's happening inside with powerful debugging and performance insights.
- [Web Dashboard](https://docs.vizra.ai/testing/web-dashboard.md): Your command center for AI agents. The Vizra ADK comes with a beautiful web dashboard that lets you monitor, test, and manage your agents - all from your browser.
- [Examples](https://docs.vizra.ai/examples/overview.md): Real-world agent examples with video walkthroughs
- [Customer Support Agent](https://docs.vizra.ai/examples/customer-support-agent.md): Build a complete customer support agent with order lookup and refund processing
- [Shopping Assistant](https://docs.vizra.ai/examples/shopping-assistant.md): Build a personal shopping assistant with cart management and preference learning
- [Artisan Commands](https://docs.vizra.ai/api-reference/artisan-commands.md): Your CLI toolkit for building amazing AI agents. Master these commands to create, debug, and deploy intelligent agents with ease.
- [Agent Class Reference](https://docs.vizra.ai/api-reference/agent-class.md): Master the heart of Vizra ADK. The BaseLlmAgent class is your gateway to building incredible AI agents. Every property, method, and secret revealed.
- [Tool Class Reference](https://docs.vizra.ai/api-reference/tool-class.md): Unleash your agent's superpowers. The ToolInterface is your gateway to giving agents incredible abilities. From database wizardry to API mastery - build tools that make agents unstoppable.
- [Workflow Class Reference](https://docs.vizra.ai/api-reference/workflow-class.md): Orchestrate agents like a conductor. The BaseWorkflowAgent turns you into a master of AI coordination - chain agents together, create complex flows, and build intelligent automation pipelines.
- [Evaluation Class Reference](https://docs.vizra.ai/api-reference/evaluation-class.md): Turn your agents into perfection machines. The BaseEvaluation class is your quality assurance superhero - test, validate, and polish your agents until they shine.
- [Events API Reference](https://docs.vizra.ai/api-reference/events.md): Become an agent lifecycle ninja. Master every moment of agent execution with comprehensive event hooks. Monitor, debug, extend - turn your agents into observable, intelligent systems.
- [Error Handling](https://docs.vizra.ai/api-reference/error-handling.md): Handle errors like a pro. The Vizra ADK provides a robust error handling system with specialized exceptions to help you build resilient AI agents that gracefully handle the unexpected.
- [Tracing Classes API Reference](https://docs.vizra.ai/api-reference/tracing-class.md): Complete API reference for the TraceSpan model and Tracer service that power ADK's tracing system.

## OpenAPI Specs

- [openapi](https://docs.vizra.ai/api-reference/openapi.json)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.