Frequently Asked Questions
Practical answers for billing, quotas, compatibility, and support workflows.
1
Billing and plan logic
Billing combines credit-based usage and package options depending on workload style.
- Use free and pay-as-you-go for gradual adoption and experiments.
- Use AI Max packages for higher dedicated usage requirements.
- Review usage history before monthly planning cycles.
- Apply budget controls per workspace for predictable spend.
2
Limits, quotas, and key behavior
Limits are enforced at key and workspace levels to protect reliability and budget.
- Per-key usage limits can stop runaway usage.
- Different plans may have different practical throughput envelopes.
- Expired or revoked keys should be rotated in applications promptly.
- Alerting should be configured before limits are reached.
3
Model compatibility and migration
Application behavior may differ across models, so migrations should be tested deliberately.
- Do not assume identical output structure across model families.
- Evaluate prompts and fallback behavior before switching defaults.
- Test safety and formatting expectations for each critical workflow.
- Roll out model changes in phases instead of global cutovers.
4
Support and escalation
Support can resolve issues faster when reports include reproducible technical context.
- Include workspace ID, timestamp, and impacted endpoint.
- Attach request IDs, status codes, and sampled payloads.
- Describe expected vs actual behavior clearly.
- Escalate outages and security incidents immediately.