OneStream XF Skill Chat — A Browser-Only Demo

A static GitHub Pages demo that answers OneStream XF questions with the public onestreamxf-skill loaded as the agent's working skill — running entirely in the browser, no backend.

Anoop Sharma

12 May 2026 — 1 min read

GitHub: anoop22/onestreamxf-chat · Live: anoop22.github.io/onestreamxf-chat

A static demo that puts the onestreamxf-skill in front of anyone who wants to try it. No install, no backend, no API key on a server — just a web page that runs an agent loop locally with the skill loaded.

How it works

GitHub Pages serves a static React/Vite app
The browser fetches the public skill Markdown files directly from GitHub
@mariozechner/pi-agent-core runs the agent loop and tool execution
@mariozechner/pi-ai streams responses from OpenRouter
The user supplies an OpenRouter API key, stored in browser localStorage only

Why a browser-only build

The whole point of publishing a skill is that other people can use it. A static demo lowers the trial cost to zero — open a URL, paste a key, ask a OneStream question. There is no server-side surface to maintain, no auth, no secrets to leak. The default model is openrouter/free so the trial is genuinely free for most users.

Local dev

npm install
npm run dev

Build with npm run build — output goes to docs/ with the Vite base path set for GitHub Pages.

Vikram AI — The Evolved Form of Vaman

Trivikrama, the evolved form of Vaman AI. The codebase Vaman is growing into — an autonomous, self-learning domain expert agent built on the same OpenClaw and pi-mono foundation.

Voice App QA Recorder — Catching Bugs Automated Tests Miss

A reusable local QA harness for voice-first and multimodal apps. Records a real browser session as synchronized screen video, microphone audio, system audio, console logs, app events, and stills — then generates a review file.

OneStream XF Domain Intelligence Skill

A modular skill package that helps AI agents answer OneStream XF questions with real domain judgement — disambiguating overloaded terminology, preserving compound concepts, and respecting architectural boundaries.

Agentic RAG Lab — A Minimal LangGraph Demo

A single-file version of LangGraph's Agentic RAG tutorial. Useful as a starting point for reading the agentic-RAG pattern end-to-end without wading through a multi-module example.