Cost Controls

Understand and control the costs of running multi-model memos.

Cost per memo

Each memo involves 5 API calls — one to each of the 4 advisors, plus one synthesis call. The cost depends on the input/output token counts:

• Simple question (~500 tokens input) — $0.05-0.10
• Detailed question with context (~2,000 tokens) — $0.15-0.25
• Complex question with briefing (~5,000 tokens) — $0.25-0.40

Reducing costs

• Use fewer providers — Configure only the advisors you need. Two advisors still produce valuable synthesis.
• Keep questions concise — Long context means more tokens. The Briefing extracts structured context efficiently.
• Use criteria — Criteria focus the advisors on what matters, reducing tangential analysis.

Monitoring usage

The Decision Memo includes metadata about the models and tokens used, so you can track costs per memo. Monitor your provider dashboards for real-time spending:

• OpenAI — platform.openai.com/usage
• Anthropic — console.anthropic.com/usage
• xAI — console.x.ai/usage
• Google — aistudio.google.com/usage

Setting budgets

All four providers support spending limits on their dashboards. Set monthly caps to avoid surprises. We recommend starting with $10-20/month per provider while testing, then adjusting based on actual usage patterns.