Stop guessing.
Run a 7-seat AI council in 30 seconds.
Three free LLMs debate every question. Premium seats join when stakes are real. One synthesized verdict — disagreements surfaced, not hidden. The §73 ladder picks cheap-first, so you pay 29× less than always-premium APIs.
5 lifetime runs free · No credit card · Cancel any time
Image shows roadmap. Live today: Gemini · Llama via Groq · GPT-OSS. Claude, Perplexity & Grok seats arrive in Wave 10+.
Built by 313 AI Agency · Powering decisions for 7 active clients
Live · the council in session
Seven AI seats. One table. Thirty seconds.
Each seat brings its own training, biases, and strengths. The curator surfaces where they converge — and where they fight. You decide.
Real council scene · avatars generated · scene refreshed each Wave
§73 Council Escalation Ladder
29× cheaper than always-premium AI
Most AI apps pay $1.50 per question. We pay $0.05. Same answer quality, smarter cost engine.
Why the math works
70% of questions resolve in Round 1 — free seats only (Llama, Gemini-Flash, DeepSeek). No premium fired.
Premium seats join only when free seats disagree, hedge, or stakes are flagged. The model picks itself.
Hard cap: never more than 2 premium models in any round. The 3rd adds <10% quality at 50% extra cost.
What you save per tier
| Tier | Runs / month | You save vs always-premium |
|---|---|---|
| Pulse | 50 | $73 / mo |
| Operator | 160 | $232 / mo |
| Smart★ | 500 | $725 / mo |
| Pro | 750 | $1,088 / mo |
| Supreme | 1,200 | $1,740 / mo |
Numbers based on average council cost of $0.052/run (29× cheaper than $1.50 always-premium baseline). Audit your own runs via the dashboard.
How it works
Sign up free
5 lifetime council runs at $0 — no card, no API keys to wrangle. 313 covers the cost on free providers so you can stress-test the real product.
Ask anything
Type your real question — a strategy decision, a campaign angle, a technical tradeoff. One question, three independent perspectives, no follow-up prompting tricks.
Get one verdict
In ~30 seconds: Gemini (fast facts) + Llama via Groq (creative) + GPT-OSS (deep reasoning) run in parallel — we surface where they agree, disagree, and converge. You decide.
Run AI Council from inside Claude Code
Add AI Council as an MCP server in ~/.claude.json. Get council.run, council.status, and council.get_memo as native tools while you code.
{
"mcpServers": {
"ai-council": {
"url": "https://aicouncil.me/api/mcp",
"headers": { "Authorization": "Bearer <your-token>" }
}
}
}Bearer token auto-generated when you sign up — free tier included. Full setup guide in your dashboard.
Get my MCP tokenBuilt to grow with you
Free tier is genuinely free — 313 covers the API cost on free providers. No credit card. Click any tier to choose your plan.
- Full 3-LLM council
- Keys included · zero setup
- + Claude Code MCP
- ~1-2 runs / day
- All keys included
- + Claude Code MCP
- ~5 runs / day · daily driver
- Projects + faster queue
- + Claude Code MCP
- Client workspaces + branded reports
- NotebookLM source packs
- + Claude Code MCP
- Sonnet std runs · GPT-5 premium · Opus on request
- Advanced memory + API access
- + Claude Code MCP
- White-label + team portals
- Auto Pilot 10× multi-step runs
- + Claude Code MCP
🔓 Cancel in 1 click · 📥 Export decision memos (paid tiers) · 🔐 Data encrypted, never logged · 💸 Founding-member pricing locks for 12 months
Start your first council in 5 minutes
No credit card. 5 lifetime runs to stress-test the real product.