Anthropic (Claude)
Use Anthropic (Claude) as a model provider for tasks you choose.
Model provider
01 · What Growth OX does with it
What Growth OX does with it
Anthropic (Claude) is a model provider that Growth OX can send tasks to. You choose which provider handles which kind of task, or keep a task on a model that runs on your own computer. Your provider's own terms apply to what you send it.
- Model profiles in configuration
- 5
- Models
- claude-fable-5-1, claude-opus-5-5, claude-sonnet-5-5, claude-haiku-5-5
02 · Connecting
Connecting
Add your own key for Anthropic (Claude) in the model settings. Growth OX does not resell model access.
03 · Permissions and data that leaves
Permissions and data that leaves
No special permission is recorded for this entry in the configuration.
What leaves your computer
- Chats (Asistan) on a cloud model (Claude Code / Codex subscription CLI): task text + inputs embedded in the prompt (up to 200 KB) and copied to input_dir: brain-context.md (cloud never gets local-only notes; secret-checked), project.json, status.md, items-*.jsonl, tracker.md, documents-index.txt or…
- Gmail: (a) mail content via `claude -p` and the claude.ai Gmail connector → Anthropic → Google; (b) metadata-scope reads
- LinkedIn, X: API clients refuse before any request (UNVERIFIED); reads happen only through `claude --chrome` in the person's Chrome (page text → Anthropic)
04 · Resources