Claude
Anthropic's models, and our default for agent work.
- made by
- Anthropic
- source
- Official Claude site
Magic Ship is one shop in Vancouver, BC, working remotely with clients worldwide. We are not a partner, reseller, or certified vendor of Anthropic - we just build with this.
What Claude is
Claude is Anthropic's family of large language models, reachable through the Anthropic API and also through Amazon Bedrock and Google Vertex AI. The line runs from small fast models to larger reasoning models, all sharing one messages API, one tool-use format, and a long context window.
How we use it
Claude is where we start for agent work: tools declared as JSON schemas, structured output validated at the boundary, and prompt caching on the system prompt and retrieved context so a long-running loop stays affordable. We pin an explicit model id per environment and treat a model upgrade like any other change - it passes the eval suite before it ships. For work inside a repo we drive it through the Claude Agent SDK rather than raw API calls.
Where it is the wrong choice
There is no open-weight Claude, so an engagement that requires inference inside the client's own network rules it out and the work goes to Llama or Mistral on vLLM. It is also not the cheap answer for high-volume classification, where a small model tuned on the client's own labels does the job for less.
Service lines it turns up in
Related tools
More in Model providers
Other tools in the same service lines
Building something on Claude?
Send the problem rather than a job spec. You get an answer on scope, on fit, and on whetherClaude is even the right call for it.
Start a projectClaude and Anthropic are trademarks of their respective owners, used here to say what we work with.