AI & ML interests
LLM observability, agent monitoring, telemetry, OpenTelemetry, production-LLM, token cost tracking, AI agents
Voight
Observability for AI agents and production LLM apps. One backend, one dashboard, one SDK family — prompts, tokens, tool calls, cache reads, latency, errors, and per-user cost attribution surfaced live at voight.xyz.
Same wire format every observability tool listed in the Vercel AI SDK docs consumes — OpenTelemetry GenAI semantic conventions. No vendor lock, no custom middleware.
SDK family
| Package | Use case |
|---|---|
@voightxyz/sdk |
Hooks for Claude Code & Cursor + library mode for autonomous agents |
@voightxyz/openai |
Wrap the OpenAI Node SDK in production apps |
@voightxyz/anthropic |
Wrap the Anthropic Node SDK in production apps |
@voightxyz/vercel-ai |
OpenTelemetry SpanExporter for the Vercel AI SDK |
Provider coverage
Pricing + telemetry capture for OpenAI (GPT-4o, o1, o3), Anthropic (Claude 4.5, Haiku 4.5, Opus 4.7), Google (Gemini 1.5 / 2.0), and Z.ai / Zhipu (GLM-5, GLM-4.7, GLM-4.5, vision, OCR). The wrappers also work zero-code against any OpenAI-compatible endpoint (Together, Groq, Fireworks, DeepInfra, Cerebras, etc.) by overriding the baseURL.
Get started
cd your-app
npx -y @voightxyz/sdk init