I built El Luchador to get page-aware AI without black-box backends. It’s a lightweight Chrome sidebar that reads only the page you’re on (no copy/paste) and applies your own prompts or presets.

What it does • Summarize/explain/analyze/translate the current page. • BYO OpenAI API key (you control spend and data). • Import a preset from any OpenAI GPT page (capture title + prompt) or create your own one-click button, then run it against the visible DOM. • Free, no account.

What’s different • Page context via activeTab → faster workflow, less overfetch. • User control over models/prompts; no vendor lock-in. • Minimal footprint (Chrome MV3, JS; content script + service worker).

Privacy & permissions No server. Everything runs in the browser except the API call you initiate with your key. Permissions: activeTab, storage.

Limitations / roadmap Chromium desktop only for now. Working on Anthropic/Gemini adapters, local embeddings for “read later,” and preset sharing. Open to security/perm audits and heavy-page edge cases.

I’d love feedback on UX, security, and performance trade-offs. Happy to answer questions and share code snippets of the prompt builder / DOM selection pipeline.