Vizra ADK
Build intelligent AI agents with Laravel's elegant syntax
FreeAI Agents4,713 impressions#3 of its week
Vizra ADK is a Laravel package for building autonomous AI agents with persistent memory and tool integration.
- for
- Laravel developers building AI-driven applications.
- pricing
- freemium
- license
- MIT
Key features
- Multi-model support — Connect agents to OpenAI, Anthropic, Google Gemini and other LLMs.
- Persistent memory — Agents retain context across runs via Laravel storage.
- Tool delegation — Agents can call and orchestrate sub-agents and custom tools.
- Livewire dashboard — Visual UI for monitoring eval runs and trends inside your app.
- Streaming responses — Receive partial LLM output as it is generated.
- Comprehensive tracing — Full logs of tool calls, arguments, costs and timings.
- Evaluation framework — Write Pest tests to evaluate agents at scale with baselines and CI gates.
- Artisan & Eloquent integration — Use familiar Laravel patterns to manage agents and eval data.
Use cases
- Validate that an AI support bot only answers from documented policies.
- Detect regressions in agent performance across CI builds.
- Track cost, token usage and tool call order for production agents.
- Share evaluation history across a team via Vizra Cloud.
Vizra ADK vs alternatives
 Vizra ADK |  LangChain |  CrewAI |  OpenAI Codex |  Claude Code | |
|---|---|---|---|---|---|
| Best for | Laravel-native AI agent framework | General-purpose LLM app framework | Multi-agent orchestration platform | Coding assistance | Agentic coding in IDEs |
| Pricing | Freemium | Free | Subscription | Subscription | Subscription |
| DevHunt upvotes | 4 | 0 | 0 | 0 | 0 |
| Launched | Aug 2025 | — | — | — | — |
- Vizra ADK vs LangChain: LangChain is a generic LLM framework, while Vizra is Laravel-specific with built-in evals and dashboard.
- Vizra ADK vs CrewAI: CrewAI focuses on multi-agent orchestration, whereas Vizra provides Laravel integration and evaluation tooling.
- Vizra ADK vs OpenAI Codex: OpenAI Codex is a coding assistant, while Vizra is a full agent framework with testing and CI support for any domain.
- Vizra ADK vs Claude Code: Claude Code targets terminal/IDE coding, whereas Vizra offers a Laravel package for building and evaluating any AI agent.
Vizra ADK FAQ
Does Vizra run my LLM model?+
No, evals run in your own app using your own model keys; Vizra never accesses your models.
What data leaves my network?+
Only scores, timings, cost, git metadata and optional per-sample detail (prompt, response, tool calls, judge reasoning).
Can I self-host the cloud component?+
The framework and local dashboard are open source and self-hosted; the cloud service cannot be self-hosted today.
Do paid tiers charge per seat?+
No, all paid plans include unlimited seats.
How does Vizra differ from Pest’s built-in evals?+
Pest’s checks are quick and unrecorded; Vizra’s toPassEval adds dataset storage, sampling, tool-call assertions and persistent baselines.
Summarized by DevHunt from vizra.ai · Oct 4, 2026. Details may change; check the official site.
.png?width=64)
