Framework Integrations
ID Wispera provides first-class integration providers for popular AI agent frameworks. Each integration wraps the vault with framework-specific credential retrieval, caching, and audit logging.Architecture
Every integration provider extends a shared base credential provider:- Credential retrieval —
get(passportName)andgetValue(passportName)to fetch credentials by passport name - Platform shortcuts —
getOpenAIKey(),getAnthropicKey(),getAzureKey()for common platforms - Auto-detect LLM config —
getLLMConfig(passportName, model?)reads the passport’s platforms and returnsapi_key+model+provider - Listing —
listAvailable(platform?)to enumerate credentials, optionally filtered - Caching — Optional in-memory cache (thread-safe via
sync.RWMutexin Go) - Audit logging — Every access is logged with actor, platform, and purpose
SDK Parity Matrix
| Integration | TypeScript | Python | Go |
|---|---|---|---|
| Base Provider | WisperaCredentialProvider | WisperaBaseProvider | CredentialProvider |
| OpenAI Agents SDK | Yes | Yes | Yes |
| Google A2A Protocol | Yes | Yes | Yes |
| Slack | Yes | Yes | Yes |
| LangChain | Yes (new) | Yes (existing) | N/A |
| CrewAI | N/A | Yes (existing) | N/A |
Base Provider
The base provider is the foundation all integrations share. You can also use it directly when you don’t need a framework-specific wrapper.Integration Guides
OpenAI Agents SDK
Agent keys, tool auth headers, and multi-agent handoff credentials.
Google A2A Protocol
Agent cards, token validation, and task-scoped credentials for agent-to-agent communication.
Slack
Bot tokens, app tokens, user tokens, webhooks, and Socket Mode support.
MCP
Connect to AI agents via the Model Context Protocol.
Next Steps
SDK usage
Core SDK operations across TypeScript, Python, and Go.
Provisioning
Programmatically create API keys at vendor APIs.
Audit log
Review credential access history from all integrations.