Open source alternative to Chess.com
Last verified
Run a self-hosted chess server with Stockfish game review, a bring-your-own-LLM AI coach, family profiles, and real-time PvP in one Docker container.
Patzer's multi-user profiles let parents track a child's progress alongside their own, with kid mode that tones down technical language and adds blunder warnings. Per-profile languages mean family members can play and get coached in their own language. The admin console controls who can register.
You only need chess for a single player with no kid-safety requirements. A local desktop app or Lichess handles that with less setup.
If you have an Ollama server on your LAN, connecting the AI coach takes one URL in Patzer's admin panel. Game Review works the same day; the coach reads from the same game history that analysis builds. No additional infrastructure is needed.
You have no local inference server and are not willing to set one up or use a cloud API key; the coaching voice will not be available.
Patzer syncs your public Chess.com and Lichess games automatically (Chess.com can sync every 15 minutes), analyzes them with Stockfish, and keeps the full review history locally. You get Chess.com-style move classifications and an eval graph without relying on a cloud service.
You need analysis depth beyond Stockfish's bundled binary, or you play on platforms other than Chess.com and Lichess.
Every analyzed game, coaching note, and rating history lives in a SQLite database inside the Docker volume on your own machine. Patzer sends nothing out except an optional six-hourly version check to GitHub, which can be disabled in Admin.
Your primary concern is playing strength and access to a large pool of online opponents. Cloud platforms like Lichess offer more players and stronger community features.
Serious chess players rely on game review and coaching to improve, but the main online platforms restrict those features behind paid tiers. Chess.com charges for full move analysis, AI coaching, and detailed lesson access. Moving to a paid tier works until pricing changes or features shift, and it gives you no control over where your game history lives.
For a household with multiple players at different skill levels, the problem compounds: family accounts cost more, kid-safe modes vary by platform, and there is no way to run everything on your own hardware with a shared leaderboard or a coach tuned to your specific game history.
Pull your public Chess.com or Lichess games (or paste a PGN) and analyze with bundled Stockfish. Every move is classified as Brilliant, Great, Best, Excellent, Good, Inaccuracy, Mistake, Miss, or Blunder. The review shows accuracy percentage, estimated Elo, an eval graph, top engine lines, key moments, and a "What's the threat?" probe that explains each tactic in plain text.
Point Patzer at an Ollama or vLLM host on your network (or a DeepSeek API key if you have no local GPU). The coach explains why a move works or fails, shows the better alternatives, and connects your mistake to patterns in your own recent game history. It never contradicts the engine, because the LLM is given Stockfish's verdict first.
Seven Stockfish tiers from Kid to maximum strength, all standard time controls, and up to six premoves queued on the board. The same server also hosts real-time PvP between profiles over WebSocket, with draw offers, takebacks, and one-click rematch.
Each user gets a profile with their own rating, language setting, audience tuning for the AI coach, and optional kid mode that uses friendly piece names like "horsey" and adds blunder warnings. The admin console controls sign-up (open, invite-only, or closed) and assigns user roles.
69 interactive lessons across four skill levels, from piece movement to tactics, mating patterns, and endgames. The opening trainer drills 16 built-in main lines or custom repertoires; moves you miss return in a daily review queue.
The UI and AI coach prompts are available in English, Bulgarian, Spanish, German, and Russian. Adding a language requires one table entry per file, documented in CONTRIBUTING.
Chess.com combines live matchmaking, game review, AI coaching, puzzles, and lessons in one cloud service. Patzer covers the analysis and coaching workflow in a container you run at home.
| Feature | Patzer | Chess.com |
|---|---|---|
| License | MIT | Proprietary |
| Self-hosting | Yes | No |
| Game review | Stockfish, bundled | Engine analysis, paid tier |
| AI coach | BYO LLM (Ollama, vLLM) | Chess.com AI, paid tier |
| Multi-user / family | Yes | Limited on free tier |
| Multilingual coaching | Yes (EN/BG/ES/DE/RU) | English interface only |
| Cost | Free to self-host | Free tier + paid plans |
| Telemetry | None | Platform data retained |
Patzer is the better fit when you want your game history on your own hardware, a family server with kid-safe profiles, or an AI coach running against a local LLM you already own. The MIT license means a chess club can run Patzer for its members with no per-user or per-server licensing cost.
Chess.com is the better fit for live matchmaking against a large player pool, real-time tournaments, and the social features of a large online chess community. Patzer's PvP is limited to profiles on the same server; playing against strangers requires Chess.com, Lichess, or a similar service. Chess.com's puzzle and lesson library is also larger than Patzer's 69 interactive lessons.
Self-host Patzer with Docker; the container includes Stockfish and starts on port 8800.
```bash
docker run -d \
-p 8800:8800 \
-v patzer-data:/app/data \
--name patzer \
ghcr.io/SikamikanikoBG/patzer:latest
```For game review and AI coaching on games you've already played, yes. Patzer imports your public Chess.com and Lichess games, analyzes them with Stockfish, and provides an AI coaching voice if you have an Ollama or vLLM server. It does not replace Chess.com's live matchmaking against strangers, its puzzle rush, or its daily puzzle stream; Patzer is designed for analysis and learning on your own recorded games.
Any model reachable through Ollama or vLLM works, including small quantized models like Llama 3.2 3B. Because Patzer's server works out all chess facts first and hands them to the model as a structured prompt, a small model is sufficient. The setup wizard lists available models once you point Patzer at your Ollama URL. DeepSeek's hosted API is also supported if you have no local GPU.
The LLM coach is separate from the playing engine. Games are played against Stockfish at one of seven strength tiers. The coach reviews moves after a game is complete, not during play. There is no mode where the LLM itself plays you.
Patzer checks GitHub every six hours and shows a notice in the admin panel when a newer release is available. To update, pull the new image and restart: docker pull ghcr.io/SikamikanikoBG/patzer:latest && docker restart patzer. If you use docker-compose with the Watchtower label, updates can be automated.
The core features work offline: playing against Stockfish, reviewing games from PGN paste, and the AI coach if your Ollama host is on the local network. The features that need internet access are importing games from Chess.com or Lichess, and the optional GitHub version check. The version check can be disabled by setting UPDATE_CHECK=0 in the container environment.
Alternative to
Chess.com
Ad-free open source chess server with tournaments and puzzles.
Alternative to
Muse
Self-hosted AI assistant for WhatsApp, Telegram, and more
Alternative to
Superhuman
AI email assistant that drafts replies and clears your inbox
Alternative to
Grok Bot
AI teammates you own: your keys, your model, your machine.
Alternative toGrok Bot
Self-hosted AI agent management with org charts and budgets
Alternative to
Aspose.Total
Office automation CLI for AI agents, no Office install required
Self-host Patzer on a Hostinger VPS
Keep it running 24/7, even when your laptop is off. A Hostinger VPS gives it a server of its own.
Choose a plan
Starts from $6.49/mo
Apply
Extra 10% off yearly plans
Install Patzer
On a server that stays on
Start building
Focus on the app, not the server
Sponsors
Alternative toSuperhuman
AI email assistant that drafts replies and clears your inbox
Alternative toClaude Managed Agents
Open-source agent deployment API. Self-host or use Omnara Cloud.
Alternative toGrok Bot
AI teammates you own: your keys, your model, your machine.
The projects we add and every new guide. Free, unsubscribe any time.
Claude Code to 
OpenCode
Jira to 
Windshift
Notion to 
AppFlowy
Slack to 
Mattermost
Linear to 
GitLab
Wispr Flow to 
Handy
Arc to 
Zen Browser
CapCut to 
Concat
Grok Bot to 
rakazo
Superhuman to 
Inbox Zero
Airtable to 
AITable
Muse to 
OpenClaw
Railway to 
Coolify
GitHub Copilot to 
Cline
Datadog to 
Netdata
Microsoft Teams to 
Rocket.ChatConfluence to 
Docmost
Pinecone to 
Qdrant
Zapier to 
n8n
Devin AI to 
OpenHands
Firebase to 
Supabase
1Password to 
KeePassXC
Contentful to 
Strapi
Adobe Premiere Pro to 
Shotcut
