Beto Muniz
Beto Muniz is an individual Brazilian-Australian software engineer and open-source creator, not a company. He builds software across product engineering, UX, developer experience, back end, and AI; his Preseason.ai project makes AI-driven developer-tool recommendations measurable, inspectable, and reproducible.
At a Glance
- Software developers and AI-assisted/vibe-coding builders
- Developer-tool vendors and platform teams
- AI/LLM researchers and technology analysts
- Mac developers and Postgres users
AI Tools by Beto Muniz
(1)
Preseason
AI Tool Recommendation Benchmark
Discussions
No discussions yet
Be the first to start a discussion about Beto Muniz
Latest News
Published 'Credentialism is coming for AI agents,' discussing agent access, trust, and possible certification or gatekeeping for agents.
Published 'Why AI agents are not more popular outside the tech industry,' analyzing slow consumer and non-technical adoption of AI agents.
Released Widen 0.1.1, adding and hardening text-to-SQL evaluation, OpenRouter transport/model pinning, privacy-gated database inspection, local-model support, and launch UX.
Announced Preseason.ai in 'Make a devtool agents want,' describing its open-source benchmark and examples for email, databases, search, and LLM observability recommendations.
Products & Services
An open-source benchmarking platform that measures which developer tools AI models recommend. It freezes prompt versions, eligible categories, model snapshots, and configuration settings, then aggregates decisions into public rankings, head-to-head comparisons, and methodology notes.
A free, open-source native macOS Postgres GUI. It introspects database schemas, drafts SQL from natural-language questions using a configured cloud model or Apple's on-device Foundation Model on eligible Macs, shows SQL for review, and runs only what the user approves.
Market Position
Preseason is an open, reproducible alternative to anecdotal developer-tool rankings: it measures what AI models actually recommend under frozen, realistic prompts rather than judging tools abstractly. Widen combines a native macOS Postgres workflow with human-reviewed SQL, local-first options, and explicit safety gates rather than autonomous execution.
Leadership
Executive Team
Beto Muniz
Creator and software engineer
Brazilian-Australian software engineer; his official bio describes prior work at Good Pair Days, Leonardo.ai, Canva, and other technology companies, and identifies him as the creator of Preseason.ai and Widen.dev.
Founding Story
Beto says he created Preseason.ai out of curiosity about how much influence current LLMs have when recommending developer tools to people building web apps with AI. He turned that question into an open-source benchmark that sends realistic prompts to a fixed panel of AI models and publishes the resulting tool recommendations.
Business Model
Revenue Model
The public projects are open-source; Preseason is presented as an open-source benchmark and Widen as a free, open-source application.
Pricing Tiers
Free, open-source native macOS Postgres GUI; cloud text-to-SQL requires a provider/API key configured by the user.
Target Markets
- Software developers and AI-assisted/vibe-coding builders
- Developer-tool vendors and platform teams
- AI/LLM researchers and technology analysts
- Mac developers and Postgres users
- Developers and teams evaluating which database, auth, hosting, payments, email, observability, AI, or other tools AI coding assistants recommend.
- Researchers, developer-tool companies, and technology observers studying AI-driven software-tool adoption and recommendation bias.
- Mac developers and database users who want local Postgres exploration and natural-language SQL drafting while retaining human review before execution.