Use sample scenarios, manual strategy briefs, and Decision Duel to compare your plan with AI advice. Experimental integrations are clearly marked.
Three dilemmas. Two minutes. One truth about you. No download needed.
Already have a profile? Try Start · See your full profile → · AI Strategy Briefs →
New here? Read the complete guide →
Playing a real strategy game right now? Start with a five-minute strategy second opinion.
3 free turn plans per day — no download, no API key.
Bring your own AI. Each mode is labeled by what it can actually do.
Choose the model used for advice in supported modes
Get an API key from your AI provider (takes 60 seconds)
Pick a labeled wired game, chat mode, or manual advisor
For advisors, describe the situation manually; wired and experimental modes show their own setup status
We're honest about what works today so you know exactly what to expect.
Each personality changes how the AI reasons. Same intelligence, different strategy.
Military rushes, dominance through force
Alliances, trades, cultural victories
Resource optimization, tech advantage
Defensive walls, retaliates only when attacked
Random betrayals, unpredictable decisions
Minimal persona flavor for comparing advice styles.
Experimental chess spectator mode for comparing models where configured. Other games are not part of the basic release-candidate scope until their adapters are verified.
Comparison surfaces stay experimental until verified
Other adapters stay gated until they pass their own release checks.
Template-based play-by-play; LLM commentary configurable in Arena setup
Experimental; not the recommended first stop for basic RC testing.
You pay for the integration layer. You bring your own AI key and pay your own API costs directly.
You pay your AI provider directly. Costs vary by model, prompt size, and mode.
1 brief turns
$0.01-0.15
1 comparison turns
$0.01-0.15
1 session turns
$0.01-0.30
gated turns
TBD
30-60 turns
$0.05-0.15
50-150 turns
$0.30-1.00
AI-GameFriend connects LLM AI models to honest game surfaces: basic chess where configured, manual strategy advisors, freeform chat games, and experiments clearly marked by status.
You paste your own AI API key (from Anthropic, OpenAI, xAI, or Google). AI-GameFriend is the interface layer for supported modes. Manual advisors use the situation you type; wired games show their own status. You control which AI you use and pay your own API costs directly.
It depends on the mode and model. Manual Strategy Brief and Decision Duel sessions are usually small API calls. Longer configured modes can cost more, and the app should show cost tracking where available.
Chess has a full legal board UI. Risk, Catan, Go, Checkers, Battleship, Connect Four, and Tic-Tac-Toe are freeform/chat modes where you describe the board. Grand strategy advisors use manual situation entry. Slay the Spire Watch is experimental and gated until live retest passes.
Baseline mode uses minimal persona flavor so you can compare advice styles without treating a model as a verified autonomous player.
Arena mode is an experimental chess spectator mode for comparing models where configured. It is not the recommended first stop for the v0.9.9 release-candidate flow.
The Founder's Pass is an optional early-supporter offer. It does not mean every experimental integration is ready today; each surface keeps its own status label.
Weekly tips, new calculator alerts, and exclusive lifetime deal offers.