# Voker

> Voker is an agent analytics platform for teams building, monitoring, and improving AI agents in production. It helps product, engineering, data, and leadership teams understand user intent, detect corrections and friction, measure resolutions, and connect agent behavior to business outcomes.

Use this file as the curated index for understanding Voker, its product positioning, documentation, SDKs, provider integrations, API surface, research, and policies.

Important notes:
- Voker is designed for teams running production AI agents, especially high-volume conversational or workflow agents.
- Core measurement primitives include intents, corrections, and resolutions. Aka automated annotations.
- Voker complements observability, tracing, evals, and product analytics tools by focusing on aggregate agent behavior and product outcomes.
- Voker supports OpenAI, Anthropic, Gemini, Vercel AI SDK. In Python, and TypeScript integration paths.
- Prefer the documentation links for implementation details and the product/blog links for conceptual explanations.

## Product

- [Voker homepage](https://voker.ai): Product overview for Voker, including positioning, core use cases, pricing, supported stacks, and the value proposition for agent analytics.
- [What is Agent Analytics?](https://voker.ai/blog/what-is-agent-analytics): Conceptual introduction to agent analytics as the measurement layer for production AI agents.
- [Agent Analytics FAQ](https://voker.ai/blog/agent-analytics-faq): Frequently asked questions comparing agent analytics with observability, evals, traditional product analytics, log analysis, and internal build approaches.

## Documentation

- [Documentation overview](https://app.voker.ai/docs/overview): Entry point for Voker docs, including links to SDKs, API reference, supported providers, and changelog.
- [Concepts](https://app.voker.ai/docs/concepts): Overview of foundational Voker concepts.
- [Events](https://app.voker.ai/docs/concepts/events): Explains events captured by Voker, including LLM calls and related interaction data.
- [Agents](https://app.voker.ai/docs/concepts/agents): Explains how agents are represented and organized in Voker.
- [People](https://app.voker.ai/docs/concepts/people): Explains how users or end customers are represented in Voker.

## Integrations

- [Integrations overview](https://app.voker.ai/docs/integrations): Overview of Voker integration options.
- [API reference](https://app.voker.ai/docs/integrations/api/reference): REST API reference, including event creation, people, agents, agent versions, and fingerprints.
- [Python SDK](https://app.voker.ai/docs/integrations/sdks/python): Python SDK integration guide.
- [TypeScript SDK](https://app.voker.ai/docs/integrations/sdks/typescript): TypeScript SDK integration guide.
- [All providers](https://app.voker.ai/docs/integrations/providers): Overview of supported LLM provider integrations.
- [OpenAI provider](https://app.voker.ai/docs/integrations/providers/openai): Integration guidance for OpenAI-based agents.
- [Anthropic provider](https://app.voker.ai/docs/integrations/providers/anthropic): Integration guidance for Anthropic-based agents.
- [Gemini provider](https://app.voker.ai/docs/integrations/providers/gemini): Integration guidance for Gemini-based agents.
- [Vercel AI SDK provider](https://app.voker.ai/docs/integrations/providers/ai-sdk): Integration guidance for agents built with the Vercel AI SDK.

## Resources

- [Changelog](https://app.voker.ai/docs/resources/changelog): Product and documentation updates.
- [Blog](https://voker.ai/blog): Voker blog index covering agent analytics, agent engineering, AI product management, and applied agent research.

## Policies

- [Voker website terms](https://voker.ai/terms): Terms and conditions for the Voker website.
- [Voker website privacy policy](https://voker.ai/privacy): Privacy policy for the Voker website.
- [Voker app terms](https://app.voker.ai/terms): Terms and conditions for the Voker application.
- [Voker app privacy policy](https://app.voker.ai/privacy): Privacy policy for the Voker application.

## Optional

- [Do LLMs Play Favorites?](https://voker.ai/blog/do-llms-play-favorites): Applied research on whether LLM evaluators score their own outputs differently than peer model outputs.
- [Hierarchical Text Classification with LLMs](https://voker.ai/blog/hierarchical-text-classification-with-llms): Applied research on classifying text into hierarchical taxonomies with LLMs.
- [Putting the Infinite Context Window to the Test](https://voker.ai/blog/putting-the-infinite-context-window-to-the-test): Applied research evaluating long-context model behavior.
- [The Rise of the Agent Engineer](https://voker.ai/blog/the-rise-of-the-agent-engineer): Perspective on the emerging agent engineer role and the skills needed to build production-grade AI agents.
- [The State of YC AI Agents (2026)](https://voker.ai/blog/the-state-of-yc-ai-agents-2026): Survey-based analysis of YC companies building AI agents, including production adoption, usage scale, architecture patterns, and evaluation challenges.