# Phrony

## Docs

- [Example: Building an embedded agent](https://docs.phrony.com/academy/example-embedded-agent.md): Create an agent in the Phrony dashboard, start runs from your code, and complete HITL user tasks over the HTTP API.
- [Academy](https://docs.phrony.com/academy/index.md): Hands-on guides for building with Phrony.
- [Agents API](https://docs.phrony.com/api-reference/agents.md): Start runs and list sessions for an agent. Paths are relative to your API base URL; send X-API-Key and Content-Type: application/json unless noted.
- [Authentication](https://docs.phrony.com/api-reference/authentication.md): X-API-Key, key format, and how Phrony scopes access for the public API.
- [File library](https://docs.phrony.com/api-reference/file-library.md): Presign, upload, finalize, and pass workspace files into runs via the public API.
- [API reference](https://docs.phrony.com/api-reference/index.md): Base URL, JSON API, and how Phrony’s public HTTP API fits with your integration.
- [Runs API](https://docs.phrony.com/api-reference/runs.md): Poll run status, read conversation, subscribe to events, and send input or session messages. Paths are relative to your API base URL; send X-API-Key and Content-Type: application/json unless noted.
- [Agent](https://docs.phrony.com/concepts/agent.md): What an agent is, how versions work, and what each configuration area controls.
- [File library](https://docs.phrony.com/concepts/file-library.md): Workspace file storage for agent runs
- [Integration](https://docs.phrony.com/concepts/integration.md): HTTP and built-in connectors whose operations become tools for your agents.
- [LLM provider](https://docs.phrony.com/concepts/llm-provider.md): A saved connection from your workspace to a language model vendor or a custom endpoint.
- [Manifest](https://docs.phrony.com/concepts/manifest.md): Declarative YAML that describes LLM providers, integrations, agents, versions, and triggers so you can export, share, and apply configuration in one step.
- [Secret](https://docs.phrony.com/concepts/secret.md): Securely stored credentials and sensitive values for your workspace.
- [Session](https://docs.phrony.com/concepts/session.md): One continuous piece of work for an agent: status, history, and controls.
- [Trigger ingress](https://docs.phrony.com/concepts/trigger-ingress.md): View delivery attempts, inspect payloads, and retry failed event or API starts from the Phrony dashboard.
- [User task](https://docs.phrony.com/concepts/user-task.md): Something a person must do so a run can continue—approvals, answers, and reviews.
- [Workspace](https://docs.phrony.com/concepts/workspace.md): An isolated scope for your organization: members, resources, and billing in Phrony.
- [Overview](https://docs.phrony.com/guard/index.md): Why agents need blocking and review, policy incidents, L1 tool-call rules, and how Guard fits into approvals and the Phrony dashboard.
- [L1 rules and evaluation](https://docs.phrony.com/guard/l1-rules.md): How tool-call rules are structured, merged, evaluated inline and asynchronously, exported in manifests, and illustrated with payment and email patterns.
- [Introduction](https://docs.phrony.com/index.md): Phrony — an execution environment for governed, agentic systems.
- [Packages](https://docs.phrony.com/packages/index.md): Official client libraries for the Phrony API.
- [TypeScript SDK](https://docs.phrony.com/packages/typescript-sdk.md): Install and use @phrony/sdk to call the Phrony public API with TypeScript: agents, runs, conversation, stream, and file uploads.
- [Agent in the loop (AITL)](https://docs.phrony.com/techniques/agent-in-the-loop.md): Sub-agents that ask the parent for input—escalation, user tasks, and how AITL fits the run tree.
- [Embedded agents](https://docs.phrony.com/techniques/embedded-agents.md): Run Phrony agents from your product over HTTP: deploy, API trigger, API key, optional workspace file uploads, and start runs.
- [Human in the loop (HITL)](https://docs.phrony.com/techniques/human-in-the-loop.md): HITL execution mode, approvals, user tasks, and how humans sit on the main run path.
- [Multi-agent systems](https://docs.phrony.com/techniques/multi-agent-systems.md): Why to compose agents, one session for the run tree, sub-agent mode, AITL, safety, model and schema guidance, troubleshooting and support, and runtime instruction addenda.
- [When to use Phrony](https://docs.phrony.com/when-to-use-phrony.md): Workflows vs agents, when a governed runtime fits, and tradeoffs of the agentic model.