One command
Add Anthropic to any Ashlr Stack project with a single command. Stack runs the auth flow, verifies the credential, and writes every secret slot into Phantom.
stack add anthropic
Or describe what you're building and let Claude pick it up via stack recommend:
stack recommend "LLM provider for a coding assistant" Auth flow
Paste a long-lived API key once. Stack verifies it against the provider's API before writing to Phantom.
How-to: Paste a secret key from https://console.anthropic.com/settings/keys.
Secret slots
stack add anthropic writes these 1 secret
slot into your Phantom vault:
ANTHROPIC_API_KEY
The values never leave Phantom in plaintext. Your .env file references slot
names, and stack exec -- <cmd> swaps them in at process-spawn time via
Phantom's local proxy.
Starter templates that include Anthropic
Apply a pre-wired stack with one command:
-
stack init --template claude-agent-openai-anthropic
Related providers in AI
- OpenAI — GPT, Realtime, embeddings. Key verified against /v1/models on paste.
- xAI — Grok + tool use. Key verified on paste.
- DeepSeek — Open-weight models. Key verified on paste.
- Replicate — Open-source model inference with per-second billing. API token stored in Phantom.
- Braintrust — LLM evals + observability + prompt playground. Verified against /v1/organization on paste.
FAQ
Do I need a Anthropic account to use it with Stack?
Yes — Stack provisions Anthropic on your behalf, but it authenticates as you. Paste a long-lived API key once. Stack verifies it against the provider's API before writing to Phantom. If you don't have a credential yet, create one at https://console.anthropic.com and paste it once.
Where does Stack store my Anthropic credentials?
In Phantom Secrets, an E2E-encrypted local vault. Stack writes the secret slot names (ANTHROPIC_API_KEY) into .stack.toml — the actual values live only in Phantom and never touch disk in plaintext.
Can I rotate or revoke this integration later?
Yes. Run `stack remove anthropic` to pull the Anthropic service back out (Phantom secrets deleted, MCP entry removed, .stack.toml cleaned up). Rotate the underlying Anthropic credentials in their dashboard — https://console.anthropic.com — and Stack's next `doctor --fix` will pick up the new values.