⚠Auto-generated documentation — This content was generated with AI assistance and is currently under review. Information may be incomplete or inaccurate. When in doubt, contact our support team.
Billing & Usage
CFFBRW uses a credit-based billing system. Every action that consumes compute or AI resources costs credits.
Plans
| Free | Pro | Team | |
|---|---|---|---|
| Credits/month | 100 | 1,000 | 5,000 |
| Workflows | 5 | 50 | Unlimited |
| Runs/month | 50 | 500 | Unlimited |
| MCP Servers | 2 | 10 | 25 |
| Cron Schedules | 1 | 10 | Unlimited |
| Plan Versioning | Last 3 | Full history | Full history |
| Pre-validation (T2) | No | Yes | Yes |
Credit Costs
| Action | Credits | Notes |
|---|---|---|
| Compile a workflow (AI call) | 10 | Flat rate. If the compiler auto-retries to fix validation errors, the retry cost is absorbed by the platform — you still pay 10 credits. |
| Compile (cached, same markdown) | 0 | Same content hash = cache hit, no AI call. |
| Pre-validation / Screen (Pro+) | 3 | AI screens your markdown for issues before compilation. |
| Run a workflow | 1 | Per run (not per step). |
| AI recovery (Tier 2) | 5 | Per recovery attempt when a step fails and AI fixes it inline. |
| ai_call step | 5 | Per AI-powered step in your workflow. |
| Cron / webhook trigger | 1 | Same as a manual run. |
| Email notification (platform) | 1 | Approval email sent via CFFBRW. BYOS email connectors are free. |
BYOA (Bring Your Own Agent) Costs
When you compile workflows via the BYOA MCP server, your own AI model does the heavy lifting. Here's what you pay vs. what's free:
| Action | Credits | Why |
|---|---|---|
validate_plan | 0 | Validation is pure computation (no AI call on our side). We want to encourage BYOA adoption. |
prevalidate_markdown | 3 | This runs our AI model to screen your markdown. |
run_workflow | 1 | Same as a dashboard run. |
| AI recovery during run | 5 | Recovery uses our AI model regardless of how the plan was compiled. |
The compilation AI cost is on your side (billed by your AI provider: OpenAI, Anthropic, etc.). The platform only charges for operations that use our AI models or compute.
Viewing Usage
Go to Usage in the sidebar to see:
- Current period usage — credits used, credits remaining, reset date
- Resource counts — active workflows, runs this month, MCP servers
- Usage history — bar chart of credit consumption over past months
- Plan details — your current tier and limits
The credit bar changes color as you approach your limit:
- Green: under 80%
- Yellow: 80-90%
- Red: over 90%
Managing Your Subscription
Go to Billing (under Settings in the sidebar) to:
- View your current plan and billing period
- Upgrade to a higher tier
- Manage payment method via the customer portal
- View invoices and payment history
Upgrades take effect immediately. The credit increase is prorated for the remaining billing period.
Tips to Optimize Credit Usage
- Use cached compilations — don't edit markdown unnecessarily; same content = 0 credits
- Test with simple workflows first — debug your markdown before running complex multi-step workflows
- Use conditional branches — skip expensive steps when they're not needed
- Leverage MCP tool caching — tool catalogs are cached, reducing redundant discovery calls
- Use BYOA for iteration — compile from your AI editor (Claude Code, Cursor) and only pay for validation (free) + runs. Your AI provider handles the compilation cost.
- Use connectors instead of raw HTTP — connector steps get auto-resolved parameter names, reducing compilation errors and retries