@jithesh_sarvin

trained a machine learning algorithm that could predict the future moves in threads. I had to do a lot of troubleshooting to fix the training process and i trained the couch model for about 5 hours

Created a backend for the website so that the users could sign in. I am changing focus to academy’s to get users. They could log in add their student and it generates reports automatically for each student. Also i have fixed a little bit of the ui to make is more clear and usable for others.

I am using vercel to host the from end but i am using a raspberry pi to run the backend locally and run a Cloudflare tunnel to connect these two.

I added a puzzle page which uses stock fish and a custom trained ml algorithm to find mate in two puzzles in complex position. It also takes in real games a input to improve its puzzles every time( try in out in voltchess.me

implemented a stockfish tab that runs in the browser of the user to reduce the servers stress

Added blogs and other text to increase seo optimization. I also integrated this new color theme which i asked ai to do it. In my opinion the color theme is mid and it took it only 10 seconds to do it which saved me a lot of work.

Devlog-5
I focused more on marking my getting getting feedback from the users and improving some small bugs and ui and by this process I was able to receive 140 unique users in 2 days and more than 50 new users in the past 12 hours( check out the website -https://www.voltchess.me)

Dev log-4
added the ability to load a chess game using chess.com api and lichess ai. And integrated stockfish in the browser to analyze the games. Create a simple algorithm to predict accuracy based on their mistakes( which are calculated by sudden movements in eval) try it out at https://www.voltchess.me

Devlog-3
I deployed this website to Vercel and did some marketing to see how many people would be willing to use my website. As per thier views , it is lower than what i expected and they never returned. Based on these comments, I improved my ui and added new features such as game review but it is still buggy so i did not publish that.

Devlog-2
I have worked to create a new side panel which is ridged for pc and is like a dropdown menu for phones. I also have routed 4 different pages analysis tab has the one with the chess board. I have designed my own logo to go on top to finish it off. I made black and gold color theme to make it feel premium

I finished building a local web dashboard that acts as a central control panel for my Hermes setup, making it possible to manage Ollama models, the Discord gateway, agent settings, system resources, maintenance tasks, and configuration files from a single interface instead of relying on terminal commands or manual file edits. The application launches through a simple startup script, runs a lightweight Python backend, and serves a browser-based UI that communicates with Hermes and Ollama through API calls.

Devlog 1

Today i was able to explore different options to create a chessboard using typescript. A really important step to take is what tools and programs i chose to use. I chose vite because of its clean ui and deployments

added a new cashing system which helps save data locally in the browser so that each person gets a individualized feed in the demo website.

devlog-4
• Crawlers: OpenAlex and arXiv jobs pull open-access papers (min 5 citations) while offsets are persisted in crawler_offsets.json so each topic resumes where it left off.
• Canonical catalog: All crawled papers land in data/papers.db, ensuring every service references the same SQLite store even when multiple components run concurrently.
• Enrichment: Each new row undergoes OpenAlex metadata backfill, DOI/arXiv linking, and topic/authorship tagging so everything downstream sees complete content.
• Metrics + embeddings: classification.metrics computes trending/hybrid scores and embeddings.pipeline generates vector representations that let the feed rank novelty and relevance.
• Purifier: db_purifier.py runs after enrichment, removing paywalled, duplicate, or incomplete papers while keeping PDF/ArXiv URLs up to date.
• Feed cache: Once the catalog is clean, FeedCache stores session rows and shown IDs so the API can quickly respond without re-running heavy scoring on every request.
• API: FastAPI’s /feed handler calls build_feed(db, refresh?, client_seen_ids) which builds a FeedContext containing seen IDs, soft/hard exclusions, and user-interest signals.
• Feed logic: The engine pulls candidates via _select_from_query, applies feedback-weighted scores, injects high jitter on refresh, and enforces seen-paper penalties so every refresh reshuffles without repeating the exact same order.
• Frontend: The YouTube-style carousel shows hero/trending/high-impact rows, pulls seen IDs from localStorage, calls api.getFeed(refresh, seenIds), and records events (click/save/dismiss) to teach the feed what must never reappear.
• Feedback loop: User events plus the refresh button feed back into feed.build_feed() via soft/hard exclusions, ensuring fresh papers are promoted while clicked/saved items stay hidden forever.

Smarter Fallback: If you have seen absolutely every paper in the catalog, the algorithm will now smoothly fall back to showing you the best papers you’ve already seen rather than giving you a completely blank screen.
Clearer UI & Reset Button: I updated the frontend so if you ever somehow run out of papers again, it will proudly tell you “You’re all caught up!” and give you a button to Clear your read history so you can start over while the crawlers hunt for more.

What it does
It’s a dashboard for everything Hermies-related:

See status — Is Ollama running? Is the Discord bot online? Which model is active?
Start/stop things — Start Ollama + Discord gateway with one click, or restart them individually
Switch models — Pick which Ollama model Hermes uses, pull new ones, delete old ones
Control resources — Limit GPU/CPU so models don’t stutter your PC (presets like Stability/Balanced/Performance, or per-model sliders)
Configure Discord — Bot token, allowed users, start/stop the gateway
Tweak agent settings — Memory, reasoning, timeouts, display options
Run maintenance — Health checks, view logs, list skills/cron jobs, backup
So instead of typing hermes gateway start or editing config files by hand, you click buttons and fill in forms.

How it works
You open it — Double-click Open Hermies.vbs (or .bat). That starts a small Python server and opens your browser.

Backend — server.py listens on port 7580 and handles API requests. core.py does the real work: talks to Ollama, runs Hermes CLI commands, reads/writes config files in %LOCALAPPDATA%\hermes.

Frontend — index.html, styles.css, and app.js render the UI. When you click something (e.g. “Start Gateway”), JavaScript sends a request like POST /api/gateway/start, the server runs the matching function in core.py, and the result is shown in the dashboard.

Data flow — The panel doesn’t replace Hermes or Ollama. It sits on top of them: it edits config.yaml and .env, calls hermes commands, and talks to Ollama’s API. Your actual AI still runs through Hermes + Ollama; the website is just the remote control.

In short: local web app → Python backend → Hermes/Ollama/config files on your machine.

I have updated the program and intereface to look better, it now has custom themes better crawler and more

This is like Netflix, but for research papers. It crawls the internet to index research papers and classifies them by topic. Then, a feed-based ML algorithm filters and displays a personalized list based on the user’s activity.