How does Orbit's pricing work?
How does Orbit's pricing work?
Which models and providers does Orbit support?
Which models and providers does Orbit support?
50+ models across OpenAI, Anthropic, Grok, Kimi, Meta, Mistral, DeepSeek, Qwen, MiniMax and more, all served through Amazon Bedrock — including Bedrock Mantle for the newest GPT-5 models. The catalogue keeps growing as new models ship.Browse the live list at tryorbit.cloud/models.
Do I need to change my code to use Orbit?
Do I need to change my code to use Orbit?
No. Point the OpenAI or Anthropic SDK you already use at Orbit’s base URL and swap in an Orbit API key. Chat completions and messages work the same, buffered or streamed.
Where do requests actually go?
Where do requests actually go?
Every request goes to Amazon Bedrock in
us-east-1 at that model’s list rate. The newest GPT-5 models use Bedrock Mantle. Orbit does not add a markup. There is no second provider and no failover.How is usage billed?
How is usage billed?
Successful requests consume organization credits based on billed input and output tokens, at Bedrock list rates. Failed provider calls are not charged. Track spend from the usage dashboard.
What happens when I run out of credits?
What happens when I run out of credits?
Requests are declined once the organization’s credit balance is exhausted, so you are never billed a surprise. Add credits by buying or upgrading a plan from the dashboard.
How do I keep my API key secure?
How do I keep my API key secure?
Orbit stores only a hash and a short preview of each key after it is shown once. Create separate keys per project or environment, and revoke any key instantly from the dashboard if it is ever exposed.
Can I use Claude Code or Cursor?
Can I use Claude Code or Cursor?
Yes. They are set up separately:
- Claude Code uses the Anthropic-compatible gateway
- Cursor uses the OpenAI-compatible base URL