# Valkyrie

> Run scalable, reproducible evaluations for AI agents.

- [Valkyrie documentation](https://docs.valkyrie.vals.ai/introduction.md): Run scalable, reproducible evaluations for AI agents.
- [Quickstart](https://docs.valkyrie.vals.ai/get-started/quickstart.md): Install Valkyrie, configure credentials, and start a benchmark run.
- [Choose a hosting mode](https://docs.valkyrie.vals.ai/get-started/hosting.md): Compare hosted and self-hosted Valkyrie deployments.
- [Configure the CLI](https://docs.valkyrie.vals.ai/get-started/configuration.md): Manage Valkyrie credentials, endpoints, and sandbox providers.
- [Configure sandbox providers](https://docs.valkyrie.vals.ai/get-started/sandbox-providers.md): Store sandbox-provider credentials and select a provider for runs.
- [Manage agents](https://docs.valkyrie.vals.ai/agents/manage-agents.md): Install, upload, list, download, and remove agent bundles.
- [Agent contract](https://docs.valkyrie.vals.ai/agents/agent-contract.md): Define how Valkyrie installs, configures, and runs an agent.
- [Start a run](https://docs.valkyrie.vals.ai/runs/start.md): Run an agent against a benchmark dataset.
- [Monitor runs](https://docs.valkyrie.vals.ai/runs/monitor.md): Stream progress, inspect status and errors, and list historical runs.
- [Manage a run](https://docs.valkyrie.vals.ai/runs/manage.md): Update concurrency, stop work, and resume or retry tasks.
- [Results and outputs](https://docs.valkyrie.vals.ai/runs/results-and-outputs.md): Download scores, task artifacts, or a specific run path.
- [Slack notifications](https://docs.valkyrie.vals.ai/runs/slack-notifications.md): Notify a Slack webhook as runs cross progress thresholds or finish.
- [Custom benchmark services](https://docs.valkyrie.vals.ai/benchmarks/custom-services.md): Connect Valkyrie to a benchmark service endpoint.
- [Benchmark authentication](https://docs.valkyrie.vals.ai/benchmarks/authentication.md): Store benchmark-service credentials and send request-specific headers.
- [Invoke a post-run Lambda](https://docs.valkyrie.vals.ai/integrations/post-run-lambda.md): Run an AWS Lambda function after a benchmark finishes.
- [Analyze runs with Docent](https://docs.valkyrie.vals.ai/integrations/docent.md): Send a finished run's agent outputs to an analyzer Lambda and open the resulting Docent reading.
- [Python SDK quickstart](https://docs.valkyrie.vals.ai/sdk/quickstart.md): Configure the async Valkyrie client and list runs from Python.
- [Manage runs with Python](https://docs.valkyrie.vals.ai/sdk/runs.md): Start, inspect, stream, recover, and download Valkyrie runs.
- [Inspect resources with Python](https://docs.valkyrie.vals.ai/sdk/resources.md): Query benchmark tasks, agent bundles, and benchmark services.
- [Handle SDK errors](https://docs.valkyrie.vals.ai/sdk/errors.md): Catch transport, API, and validation failures from the Python SDK.
- [CLI reference](https://docs.valkyrie.vals.ai/reference/cli/index.md): Complete generated reference for every Valkyrie CLI command.
- [valkyrie run](https://docs.valkyrie.vals.ai/reference/cli/run.md): Complete generated reference for every `valkyrie run` command.
- [valkyrie agent](https://docs.valkyrie.vals.ai/reference/cli/agent.md): Complete generated reference for every `valkyrie agent` command.
- [valkyrie benchmark](https://docs.valkyrie.vals.ai/reference/cli/benchmark.md): Complete generated reference for every `valkyrie benchmark` command.
- [valkyrie config](https://docs.valkyrie.vals.ai/reference/cli/config.md): Complete generated reference for every `valkyrie config` command.
- [Python SDK reference](https://docs.valkyrie.vals.ai/reference/sdk/index.md): Complete generated reference for the public Valkyrie Python SDK.
- [ValkyrieClient](https://docs.valkyrie.vals.ai/reference/sdk/client.md): Construct and close the async Valkyrie SDK client.
- [Runs resource](https://docs.valkyrie.vals.ai/reference/sdk/runs.md): Async operations for the Valkyrie run lifecycle.
- [Benchmarks resource](https://docs.valkyrie.vals.ai/reference/sdk/benchmarks.md): Async operations for hosted benchmark and task inspection.
- [Agents resource](https://docs.valkyrie.vals.ai/reference/sdk/agents.md): Async operations for agents uploaded to the configured tenant.
- [Benchmark services resource](https://docs.valkyrie.vals.ai/reference/sdk/services.md): Async operations for hosted and custom benchmark services.
- [Python SDK types](https://docs.valkyrie.vals.ai/reference/sdk/models/index.md): Public models and enums exported by the Valkyrie Python SDK.
- [Agents types](https://docs.valkyrie.vals.ai/reference/sdk/models/agents.md): Public agents models and enums exported by the Valkyrie Python SDK.
- [Runs types](https://docs.valkyrie.vals.ai/reference/sdk/models/runs.md): Public runs models and enums exported by the Valkyrie Python SDK.
- [Benchmarks types](https://docs.valkyrie.vals.ai/reference/sdk/models/benchmarks.md): Public benchmarks models and enums exported by the Valkyrie Python SDK.
- [Services types](https://docs.valkyrie.vals.ai/reference/sdk/models/services.md): Public services models and enums exported by the Valkyrie Python SDK.
- [Config types](https://docs.valkyrie.vals.ai/reference/sdk/models/config.md): Public config models and enums exported by the Valkyrie Python SDK.
- [Python SDK errors](https://docs.valkyrie.vals.ai/reference/sdk/errors.md): Generated exception hierarchy for the public Valkyrie SDK.
- [Self-hosting infrastructure](https://docs.valkyrie.vals.ai/self-hosting/infrastructure.md): Review the AWS architecture and deployment boundaries for self-hosted Valkyrie.
- [AWS permissions](https://docs.valkyrie.vals.ai/self-hosting/aws-permissions.md): Scope the AWS credentials used by Valkyrie runs.