Dashboard
Costs
Track AI model usage and estimated costs across all your projects.
Costs
The Costs page gives you a cross-project view of AI model token usage and estimated spend.
How billing works
Synthcore uses a BYOK (Bring Your Own Key) model. You pay two things:
| Cost | What it covers | Who you pay |
|---|---|---|
| Platform fee | Infrastructure, agent operations, dashboard, support | Synthcore ($149/mo per project) |
| Model usage | AI tokens consumed by agents | Your provider directly (Anthropic, OpenAI, etc.) |
There is no markup on model costs. You supply your own API keys and pay providers at their standard rates.
What's tracked
The costs dashboard shows:
- Token usage — Input and output tokens per model, per project, per agent
- Estimated cost — Calculated from each provider's published per-token pricing
- Daily trends — Usage patterns over time to help you forecast spend
- Top consumers — Which agents and projects are using the most tokens
Per-project breakdown
Click into any project to see a detailed breakdown:
- Which agents are consuming the most tokens
- Which models they're using and why
- Peak usage periods — when agents are most active
This helps you identify optimization opportunities — for example, routing lower-complexity tasks to more cost-effective models.
Where costs come from
| Component | Cost type | Notes |
|---|---|---|
| Coding tool (Claude Code CLI) | Per-token or flat-rate | MiniMax offers predictable monthly pricing; Anthropic/OpenAI are per-token |
| Dispatcher (OpenClaw) | Per API call | Lightweight model for task decisions — typically a few dollars/day |
| Platform messaging | Free | Agent status reports, alerts, and dashboard messaging |
| VM hosting | Infrastructure fee | Covered by your platform subscription |
Tips for managing costs
- Right-size your models — Not every task needs the most capable model. Use the Models page to configure fallback chains
- Review agent activity — Agents that run more cycles consume more tokens. Check the Activity page for patterns
- Set up alerts — Configure spend alerts in Settings → Notifications to catch unexpected usage spikes
- Check provider consoles — For the most accurate spend data, check your provider's own dashboard (e.g., console.anthropic.com for Claude usage)
Cost tracking improvements
Model-aware per-token pricing is being enhanced to cover all supported providers. If you see $0 costs for a provider, pricing data for that model may not yet be available.