AI Vendor Directory · Foundation model APIs
Will Anthropic Claude API sign a BAA?
Per Anthropic Claude API's published pages as of 2026-07-29: HIPAA BAA available (HIPAA-ready configuration). Check which plan or configuration the BAA covers before any PHI is sent.
What is Anthropic Claude API?
Anthropic's first-party API for Claude models. SMBs use it for chat features, document analysis, coding agents, and structured extraction inside their own systems.
Key facts: BAA, retention and training
- Vendor: Anthropic — https://www.anthropic.com/api
- Signs a HIPAA BAA: BAA available (per vendor) (vendor wording: HIPAA BAA available (HIPAA-ready configuration))
- Source for the BAA answer: https://trust.anthropic.com/
- Data retention: Unknown — not verified here
- Trains on customer data by default: no (all commercial services, incl. API)
- Source for the training answer: https://www.anthropic.com/legal/commercial-terms
- Last verified: 2026-07-29
Sending PHI to Anthropic Claude API? What a hospital reviewer will ask
- If your product sends protected health information to Anthropic Claude API, Anthropic Claude API is your subcontractor under HIPAA, so you need a signed BAA with it before any PHI flows, on the plan or configuration the BAA covers.
- Reviewers ask where the data is stored, how long the vendor keeps it, and whether it trains on it. Keep the vendor’s answers (and the signed BAA) with your vendor records.
Does Anthropic Claude API train on your data?
No — does not train on your data by default. Vendor-stated default: no (all commercial services, incl. API). Anthropic's Commercial Terms of Service state flatly: 'Anthropic may not train models on Customer Content from Services' — a contractual prohibition with no opt-in for standard commercial use. Customers retain rights to inputs and own outputs. Consumer Claude (Free/Pro/Max) is governed separately: a user-controlled Model Improvement setting determines whether chats/coding sessions train Claude (with up to 5-year retention for allowed data); commercial API traffic is unaffected by that setting.
What certifications does Anthropic Claude API hold?
- SOC 2 Type I & Type II
- ISO 27001:2022
- ISO/IEC 42001:2023
- HIPAA BAA available (HIPAA-ready configuration)
Security, DPA, and subprocessor links
- Security / trust page: https://trust.anthropic.com/
- Data processing agreement (DPA): https://www.anthropic.com/legal/data-processing-addendum
- Subprocessor list: https://trust.anthropic.com/
Optional, if you sell into the EU: your EU AI Act role when you build on Anthropic Claude API
Building on the Claude API typically makes you the provider of your downstream AI feature under the EU AI Act, carrying Article 50 transparency duties for your users; Anthropic carries the GPAI model-provider obligations.
What to record in your vendor or AI system inventory
- Provider: Anthropic
- Model type: third party api
- Data typically flowing to the vendor: prompts, completions, uploaded documents, tool-use payloads
Sources
- https://www.anthropic.com/legal/commercial-terms
- https://support.claude.com/en/articles/10015870-what-certifications-has-anthropic-obtained
- https://privacy.claude.com/en/articles/10023580-is-my-data-used-for-model-training
- https://trust.anthropic.com/
Disclaimer
Curated from vendor-published pages on the date above. Vendor terms change - verify against the cited sources before relying on a profile for a procurement decision. Last verified 2026-07-29. This directory is not legal advice and does not confirm any vendor’s HIPAA compliance; a BAA is between you and the vendor.
Frequently asked questions
Will Anthropic Claude API sign a BAA?
Per Anthropic Claude API's published pages as of 2026-07-29: HIPAA BAA available (HIPAA-ready configuration). Check which plan or configuration the BAA covers before any PHI is sent.
Does Anthropic Claude API train AI models on your data?
No — does not train on your data by default. Vendor-stated default: no (all commercial services, incl. API). Anthropic's Commercial Terms of Service state flatly: 'Anthropic may not train models on Customer Content from Services' — a contractual prohibition with no opt-in for standard commercial use. Customers retain rights to inputs and own outputs. Consumer Claude (Free/Pro/Max) is governed separately: a user-controlled Model Improvement setting determines whether chats/coding sessions train Claude (with up to 5-year retention for allowed data); commercial API traffic is unaffected by that setting.
How long does Anthropic Claude API keep your data?
Unknown. This directory has not verified Anthropic Claude API's retention terms; check its DPA or BAA.
What certifications does Anthropic Claude API hold?
Per the vendor's published pages as of 2026-07-29: SOC 2 Type I & Type II; ISO 27001:2022; ISO/IEC 42001:2023; HIPAA BAA available (HIPAA-ready configuration).
Selling into the EU: what is your EU AI Act role when you build on Anthropic Claude API?
Building on the Claude API typically makes you the provider of your downstream AI feature under the EU AI Act, carrying Article 50 transparency duties for your users; Anthropic carries the GPAI model-provider obligations.