I am using MVP Cafe as a live experiment: can AI agents continuously turn briefs, constraints, QA gates, and human product review into useful browser-native tools?
This is not a product launch and it is not the MVP Cafe service offer. It is a capability lab: a public place to test how far autonomous agents can go when the surrounding product system gets sharper after every build.
The question is simple: can agents continuously build useful browser tools if the briefs, constraints, review loops, and QA gates improve with them?
The workspace grows because every failure becomes a rule for the next build.
Define the tool, audience, constraints, and expected workflow.
Agents turn the brief into a browser-native single-page tool.
Shared toolbar, routing, storage, and mobile behavior are enforced.
Build, console, overflow, and mobile checks run before review.
Human product judgment catches usability issues automation misses.
Fixes become instructions for the next agent build.
The current workspace has 80 tools. The number is a live stat, not the story.
Validate, scope, roadmap, ship.
Hunt, interview, negotiate, decide.
Focus, writing, code, everyday utilities.
Search, filter, open recent tools, and jump into the latest experiments.
Explore the workspace →These are useful tools, but they also expose what agent-built products need: touch polish, layout rules, persistence, and review.
Agents can create breadth quickly. Product quality still needs constraints, QA, and human taste.
Some tools loaded on mobile but opened with cramped panels, clipped actions, or unusable workspaces.
A page can pass build and console checks while still feeling awkward under real thumbs.
Without a toolbar, storage, routing, and mobile contract, every new tool invents its own mess.
Fixes are only valuable if they become rules for the next agent build.
The first eight tools were a forcing function: small, self-contained, free browser tools shipped in public. They proved the format worked.
The newer question is bigger: can an agent pipeline keep producing tools while the product system around it keeps learning? That is why this lives inside MVP Cafe as a Lab, not as a one-off tools launch.
The old tools are still here as the archive of where it began. The workspace is where the ongoing experiment continues.
The original free tools remain available while the agent-built workspace keeps expanding.
Validate before you build
Score your startup idea across 8 dimensions — market, timing, defensibility, and more. Reality check in 5 minutes.
Track your hunt
Kanban-style job application tracker. Move applications through stages, never lose a lead.
Compare job offers intelligently
Side-by-side offer comparison. Weighs salary, equity, growth, culture. See which offer actually wins.
Your networking CRM
Track everyone you meet. Relationship strength, last contact, follow-ups. Stop losing connections.
Professional invoices in seconds
Clean, professional invoices. Line items, taxes, notes. Download as PDF. No signup needed.
Build in public tracker
Log what you ship, every day. Streaks, milestones, progress. Your building-in-public companion.
Plan, ship, grow
Visual content calendar. Drag-and-drop scheduling, platform tags, idea backlog. Stay consistent.
PM interview prep & practice
48 curated PM interview questions. Timed practice mode. STAR answer builder. Ace the interview.
Some tools are polished. Some are deliberately rough. The point is to make the agent-building process visible.
Open the current workspace →