Aira

SDKs & Integrations

Aira SDKs for Python and TypeScript, plus framework integrations for LangChain, CrewAI, OpenAI, Vercel AI, and more.

Both SDKs now use the two-step flow. Call authorize() before the agent acts. Call notarize() after the agent executes. See the Quickstart for the end-to-end pattern.

SDKs

PythonTypeScript
Installpip install aira-sdknpm install aira-sdk
PackagePyPInpm
SourceGitHubGitHub

Both SDKs have full API coverage. Pick your language:


Integration Matrix

IntegrationPythonTypeScript
LangChainpip install aira-sdk[langchain]import from "aira-sdk/extras/langchain"
CrewAIpip install aira-sdk[crewai]
OpenAI Agentspip install aira-sdk[openai-agents]import from "aira-sdk/extras/openai-agents"
Vercel AIimport from "aira-sdk/extras/vercel-ai"
Google ADKpip install aira-sdk[google-adk]
AWS Bedrockpip install aira-sdk[bedrock]
MCP Serverpip install aira-sdk[mcp]import from "aira-sdk/extras/mcp"
CLIpip install aira-sdk[cli]
WebhooksBuilt-inBuilt-in
Offline ModeBuilt-inBuilt-in
Session ContextBuilt-inBuilt-in

Trust Layer

Both SDKs include full trust layer support: W3C DIDs, Verifiable Credentials, mutual notarization, and reputation scoring.

FeaturePythonTypeScript
DID Identity (get_agent_did / getAgentDid)YesYes
Key Rotation (rotate_agent_keys / rotateAgentKeys)YesYes
Verifiable Credentials (get_agent_credential / getAgentCredential)YesYes
Credential Verification (verify_credential / verifyCredential)YesYes
Credential Revocation (revoke_credential / revokeCredential)YesYes
Mutual Notarization (request_mutual_sign / requestMutualSign)YesYes
Reputation Score (get_reputation / getReputation)YesYes
Endpoint Policy (set_endpoint_policy / setEndpointPolicy)YesYes
DID Resolution (resolve_did / resolveDid)YesYes
Trust Check (check_trust / checkTrust)YesYes
Trust Bundle (get_trust_bundle / getTrustBundle)YesYes
Trust Policy in Integrationstrust_policy={}trustPolicy: {}

On this page