Chilla is hosted software automation. Guardrails exist to keep work inside approved product boundaries, protect service integrity, and limit activity when required conditions are not met.
1. Why guardrails exist
Chilla works in a high-stakes environment. Product controls are intended to prevent unauthorized expansion of live work, limit activity when important conditions are not satisfied, and protect users and the Services.
Guardrails can block, pause, skip, or limit activity. They do not promise a profitable outcome or remove market risk.
2. Work Capacity
Work Capacity represents the service capacity available for Chilla to keep working during the applicable usage window. The user-facing product shows available capacity as a percentage.
Capacity is a service-resource control, not a financial result, performance score, speed promise, or market signal. The product may reduce available capacity as work is carried and restore it according to the applicable tier rules shown in Chilla.
3. Capacity tiers
Your subscription tier determines the amount of Work Capacity available and may also determine how many live accounts Chilla can work on.
- Chill Architect: Limited capacity.
- Chill Mode: Standard capacity.
- Deep Chill: Expanded capacity.
- Peak Chill: Highest capacity.
Current pricing and live-account allowances are listed on Pricing.
If available capacity reaches its limit, work that depends on that capacity may pause until capacity becomes available again.
4. Authorization is separate from intelligence
Chilla may understand a goal, explain available product options, or turn the material choices you make into bounded work for review. None of those things gives it unlimited authority to begin or materially expand live financial work by itself.
Where user approval is required, the material work must be reviewed and authorized before it becomes active. Material changes that require new authority return to the user.
Intelligence is not authority. Conversation does not substitute for the user authorization required by the product.
5. Per-activity limits and other safeguards
The amount the user allows Chilla to use per activity and other product safeguards apply to authorized work. Depending on the account, provider, market and work involved, Chilla may wait, skip, pause, refuse or limit an action when required conditions are not met.
These controls can reduce avoidable product risk but cannot guarantee a specific loss limit in every market or provider condition.
6. Report Back
Chilla reports back on the work in plain language, including meaningful entries, exits, skips, warnings, changes and other activity that affected the goal.
Report Back is designed to explain the story of the work without requiring you to operate a separate activity dashboard.
7. Temporary service protection
Beaverly may temporarily restrict, pause, or refuse some activity when needed to protect users or the Services, including where:
- capacity or service limits are reached;
- required provider or account information is unavailable or unreliable;
- market or system data is not suitable for safe operation;
- security, abuse, fraud, or integrity concerns are detected;
- a technical problem requires recovery before work can continue;
- legal or regulatory obligations require a restriction.
Where work cannot safely continue, Chilla may wait, pause, skip, refuse, or return to the user rather than improvising.
8. What capacity and guardrails do not mean
- Capacity does not estimate profit, quality, or expected market activity.
- A blocked or skipped action does not prove that the action would have made or lost money.
- A safeguard does not guarantee that loss cannot occur.
- More capacity does not mean better financial outcomes.
- A higher tier does not change the non-advisory, non-custodial nature of the Services.
Questions about capacity or product limits can be sent to support@beaverlyai.com.