Aten Security builds post-approval governance for AI agents and enterprise SaaS integrations. This organization publishes our open-source SDKs and developer tools.

Instrument any AI agent with policy enforcement, behavioral monitoring, and step-up authentication in minutes. Works with Claude, OpenAI, LangChain, and more.

client, _ := thoth.NewClient(thoth.Config{
    TenantID: "acme-corp",
    AgentID:  "invoice-processor-v1",
})
defer client.Close()

search := client.WrapToolFunc("search_docs", mySearchFn)
// governance runs automatically on every call
result, err := search(ctx, map[string]any{"query": "access policy"})

import thoth

wrapped = thoth.instrument_anthropic(
    {"search_docs": my_search_fn},
    agent_id="invoice-processor-v1",
    approved_scope=["search_docs"],
    tenant_id="acme-corp",
)

# Drop into your Anthropic agentic loop
result = wrapped["search_docs"](block.input)

import { wrapOpenAITools } from "@atensec/thoth";

const tools = wrapOpenAITools(
  { search_docs: mySearchFn },
  { agentId: "invoice-processor-v1", approvedScope: ["search_docs"], tenantId: "acme-corp" },
);

// Drop into your OpenAI tool-calling loop
const result = await tools[call.function.name](args);

• Docs: docs.aten.security
• Website: aten.security
• Contact: sdk@aten.security
• Security issues: security@aten.security (do not open public issues)

Thoth SDKs are Apache 2.0 licensed.