# Docent

## Docs

- [Docent Agent](https://docs.transluce.org/analysis/docent-agent.md): Ask questions about your agent runs in natural language and review a verifiable report.
- [Structured Queries (DQL)](https://docs.transluce.org/analysis/dql.md): A read-only SQL subset for structured, quantitative questions about your data
- [Exporting Data](https://docs.transluce.org/analysis/exporting.md): Export transcripts and agent run metadata for local analysis
- [Labeling Agent Runs](https://docs.transluce.org/analysis/labeling.md): Annotate agent runs with structured data
- [Analysis Overview](https://docs.transluce.org/analysis/overview.md): Use Docent to surface and measure unexpected behaviors
- [Rubrics and Judges](https://docs.transluce.org/analysis/rubrics.md): Define evaluation criteria and run LLM-based judges on agent runs
- [Rubric Refinement](https://docs.transluce.org/analysis/rubrics/refinement.md): Quantifying behaviors that require judgment to detect
- [Search and Clustering](https://docs.transluce.org/analysis/search-and-clustering.md): Search for specific behaviors in agent runs and cluster results
- [Agent Run](https://docs.transluce.org/concepts/agent-run.md)
- [Chat Messages](https://docs.transluce.org/concepts/chat-messages.md)
- [Collection](https://docs.transluce.org/concepts/collection.md)
- [LLM Output](https://docs.transluce.org/concepts/llm-output.md)
- [Metadata](https://docs.transluce.org/concepts/metadata.md)
- [Data Models Overview](https://docs.transluce.org/concepts/overview.md)
- [Transcript](https://docs.transluce.org/concepts/transcript.md)
- [Agentic Ingestion](https://docs.transluce.org/ingestion/docent-agent.md): How the Docent Agent generates ingestion scripts
- [Harbor](https://docs.transluce.org/ingestion/integrations/harbor.md): Ingest Harbor/ATIF trajectories into Docent
- [Inspect](https://docs.transluce.org/ingestion/integrations/inspect.md): Ingest Inspect logs into Docent
- [NeMo Gym](https://docs.transluce.org/ingestion/integrations/nemogym.md): Ingest NeMo Gym rollouts into Docent
- [Ingestion Overview](https://docs.transluce.org/ingestion/overview.md): Upload your data to Docent
- [Ingest via SDK](https://docs.transluce.org/ingestion/sdk.md): Ingest agent runs into Docent using the Python SDK
- [Logging Agent Runs](https://docs.transluce.org/ingestion/sdk/logging.md): Non-blocking uploads of AgentRun objects
- [Tracing](https://docs.transluce.org/ingestion/tracing.md): Automatically capture LLM interactions with Docent's tracing system
- [Welcome to Docent](https://docs.transluce.org/introduction.md): Understand and improve your benchmark scores
- [Quickstart](https://docs.transluce.org/quickstart.md): Welcome! Run your first analysis in Docent
- [Agent Run Metadata](https://docs.transluce.org/sdk/agent-runs/metadata.md): Read and write metadata on agent runs and transcript groups
- [Query Agent Runs](https://docs.transluce.org/sdk/agent-runs/query.md): Retrieve and search agent runs in a collection
- [Authentication](https://docs.transluce.org/sdk/authentication.md): Configure API keys and connection settings for the Docent SDK
- [Client](https://docs.transluce.org/sdk/client.md): Initialize and configure the Docent client
- [Manage Collections](https://docs.transluce.org/sdk/collections/manage.md): Create, list, update, and delete collections
- [Collection Metadata](https://docs.transluce.org/sdk/collections/metadata.md): Read and write metadata on collections
- [Execute DQL Queries](https://docs.transluce.org/sdk/dql/execute.md): Run queries against your data with the Docent Query Language
- [DQL Schema](https://docs.transluce.org/sdk/dql/schema.md): Introspect available tables and columns for DQL queries
- [Error Handling](https://docs.transluce.org/sdk/errors.md): Handle errors from the Docent SDK
- [Labels, Tags & Comments](https://docs.transluce.org/sdk/feedback/labels.md): Annotate agent runs with structured labels, tags, and comments
- [Ingestion Job Management](https://docs.transluce.org/sdk/jobs.md): Track the status of background processing jobs
- [SDK Overview](https://docs.transluce.org/sdk/overview.md): Install and get started with the Docent Python SDK
- [Permissions & Sharing](https://docs.transluce.org/sdk/permissions.md): Share collections with users, organizations, and the public
- [Clustering](https://docs.transluce.org/sdk/rubrics/clustering.md): Access clustering results for rubric evaluations
- [Run Evaluations](https://docs.transluce.org/sdk/rubrics/evaluate.md): Start evaluation jobs and track their progress
- [Manage Rubrics](https://docs.transluce.org/sdk/rubrics/manage.md): Create, retrieve, and list rubrics and judges
- [Browser Integration](https://docs.transluce.org/sdk/ui.md): Open collections, rubrics, and results in the Docent web UI
- [Contact Us](https://docs.transluce.org/support.md): We'd love to hear from you.

## OpenAPI Specs

- [openapi](https://docs.transluce.org/api-reference/openapi.json)