1. Get your organization ID
- Sign in at app.agnost.ai.
- Open Settings > Organization. Copy your organization ID: it’s a UUID.
org_id or AGNOST_ORG_ID (also called write_key in some SDKs). Treat it as a public identifier: it routes telemetry to your workspace, but it does not grant dashboard read access on its own.
If your conversations can contain personal or sensitive data, read Data Governance before sending production traffic. Agnost does not currently provide automatic PII redaction before ingestion, so redact or pseudonymize sensitive fields in your application first.
2. Proceed with skill-based onboarding
Install the skill in your agent environment:Integration guides
Use these guides if you want to wire Agnost yourself or review what the skill will detect.AI conversations
Track LLM calls and agent turns directly from Python or TypeScript application code.
OpenTelemetry
Forward traces from OTel-enabled frameworks such as Vercel AI SDK, Mastra, LangChain, OpenAI, etc.
MCP servers
Track tool, resource, and prompt calls from a dedicated MCP server.
MCP Toolbox
Send telemetry from Google’s MCP Toolbox for Databases.
Verify
Open app.agnost.ai and check:- Conversations for SDK-tracked interactions
- Tools for MCP tool invocations
- Raw logs for the unfiltered event stream
Next steps
- SDK index: published packages and imports.
- Authentication: API keys and JWT for dashboard APIs.
- Data Governance: production data handling.
