# Three.dev Documentation

## Round3 Documentation

- [Introduction](https://docs.three.dev/getting-started/readme.md)
- [Onboard with an AI agent](https://docs.three.dev/getting-started/onboard-with-an-ai-agent.md): Install the three.dev agent skills in your coding agent and let it wire the proxy into your codebase.
- [Glossary / Core Concepts](https://docs.three.dev/getting-started/glossary-and-core-concepts.md): Core terminology for three.dev — use cases, requests, sessions, live scoring, scores, and failure modes.
- [Planning your integration](https://docs.three.dev/getting-started/planning-your-integration.md): Decisions to make before wiring in the proxy — how many use cases to create, whether environments need separate keys, what a session is for your feature, and which tags to send.
- [Quickstart: Record your requests](https://docs.three.dev/getting-started/quickstart-record-requests.md): Set up the three.dev proxy to record your LLM requests for observability.
- [Overview](https://docs.three.dev/sending-requests/sending-requests.md): Route your LLM requests through the three.dev proxy for observability.
- [AI Providers Integration](https://docs.three.dev/sending-requests/ai-providers-integration.md): Configure the three.dev proxy for OpenAI, Anthropic, Gemini, Azure OpenAI, OpenRouter, Concentrate, Amazon Bedrock, LiteLLM, the Vercel AI SDK, and LangChain, with code examples per provider and langu
- [Supported paths](https://docs.three.dev/sending-requests/supported-paths.md): Advanced observability for the Chat Completions, Messages, and Responses APIs and Gemini's generateContent API, session grouping, and request- and session-level tags.
- [Overview](https://docs.three.dev/exploring-your-data/exploring-your-data.md): The Overview page — quality, latency, and cost for a use case over the last seven days, plus the recurring failure modes in live traffic.
- [Sessions and requests](https://docs.three.dev/exploring-your-data/sessions-and-requests.md): Read individual LLM calls on the Requests page and whole conversations on the Sessions page.
- [Filtering](https://docs.three.dev/exploring-your-data/filtering.md): Narrow the Requests and Sessions lists with the filter rail, or type filter expressions in the filter bar — two views over the same filter.
- [Finding your way around](https://docs.three.dev/exploring-your-data/navigating.md): Jump to any use case or section with Search.
- [Chat](https://docs.three.dev/exploring-your-data/chat.md): Ask the assistant about your recorded traffic, then continue the conversation in the AI agent you use on your code.
- [Overview](https://docs.three.dev/live-scoring/live-scoring.md): Score a sample of production requests pass/fail, so quality on live traffic is measured continuously.
- [Failure modes](https://docs.three.dev/live-scoring/failure-modes.md): three.dev groups live scoring failures into labeled failure modes, so recurring problems in production traffic read as a short list instead of individual requests.
- [Overview](https://docs.three.dev/for-ai-agents/for-ai-agents.md): Machine-readable entry points for AI agents consuming three.dev documentation.