For AI agents & developers

How agents can use IdentityConnectors

IdentityConnectors is an evidence-first discovery layer for identity, authentication, authorization, credential and workload-identity infrastructure.

1. Describe the task

Express delegated access, service-to-service identity, enterprise SSO, fine-grained authorization, MCP, workload identity or credential requirements in natural language.

2. Query the recommender

https://identityconnectors.com/api/recommend?q=... returns the deterministic shortlist. https://identityconnectors.com/api/ai-recommend?q=... adds an AI comparison constrained to those same verified candidates.

3. Inspect evidence semantics

Unknown protocol support stays unknown. The recommender never upgrades an Unknown field to supported.

4. Follow provider evidence

Use provider and documentation URLs to verify requirements before implementation.

Example agent queries

My agent needs Gmail and Salesforce access on behalf of users, with OAuth tokens kept outside the model and MCP support.

Find identity infrastructure for delegated Microsoft 365 access plus workload identity.

I need fine-grained authorization for a RAG agent so users only retrieve documents they are allowed to see.

Which identity systems explicitly support AI-agent identity or MCP authorization?

Machine entry points

Catalog: https://identityconnectors.com/api/catalog.json
Structured recommendations: https://identityconnectors.com/api/recommend?q=...
AI comparison: https://identityconnectors.com/api/ai-recommend?q=...
ARD: https://identityconnectors.com/.well-known/ard.json
llms.txt: https://identityconnectors.com/llms.txt