Documentation
Documentation
{"content":"Valdr gives you structured, scored, and auditable control over AI agent workflows. Whether you're setting up for the first time, exploring the UI, or building reusable agent packs — start here.\n\n{{\u003c cards \u003e}}\n {{\u003c card link=\"/valdr/docs/features/\" title=\"Features\" subtitle=\"Product guides for Workspace Knowledge, agent sessions, reviews, and more\" icon=\"sparkles\" \u003e}}\n {{\u003c card link=\"/valdr/docs/workflows/\" title=\"Valdr Workflows\" subtitle=\"Build and operate durable, reviewable agent processes\" icon=\"arrows-expand\" \u003e}}\n {{\u003c card link=\"/valdr/docs/getting-started/\" title=\"Getting Started\" subtitle=\"Install, configure, and start using Valdr\" icon=\"play\" \u003e}}\n {{\u003c card link=\"/valdr/docs/ui/\" title=\"Valdr UI\" subtitle=\"Navigate the web console: dashboard, plans, tasks, agents\" icon=\"home\" \u003e}}\n {{\u003c card link=\"/valdr/docs/valdr-mcp/\" title=\"Valdr MCP\" subtitle=\"The complete MCP tool surface your agents can call\" icon=\"code\" \u003e}}\n {{\u003c card link=\"/valdr/docs/valdr-packs/\" title=\"Agent Packs\" subtitle=\"Bundle charters, prompts, and capabilities for reuse\" icon=\"user-group\" \u003e}}\n {{\u003c card link=\"/valdr/docs/valdr-provider-packs/\" title=\"Providers \u0026 Presets\" subtitle=\"JSON configuration for AI providers and agent provider presets\" icon=\"cube\" \u003e}}\n {{\u003c card link=\"/valdr/docs/troubleshooting/\" title=\"Troubleshooting\" subtitle=\"Fix install, license, MCP, provider, session, and pack issues\" icon=\"support\" \u003e}}\n{{\u003c /cards \u003e}}\n\n## Build with Valdr\n\nChoose the interface for the work you want to automate:\n\n| I want to… | Start here | Check before running |\n| --- | --- | --- |\n| Build a process with steps, waits, and approvals | [Definition Authoring](/valdr/docs/workflows/definition-authoring/) | Validate the definition, inspect warnings, then test with safe inputs |\n| Turn a script or CLI into a reusable action | [Build Your First Tool](/valdr/docs/workflows/user-tools/quickstart/) | Validate the manifest and JSON contract, package it, and review import preflight |\n| Let an agent operate my local workspace | [Valdr MCP](/valdr/docs/valdr-mcp/) | Connect to the same data directory, discover tools, and read each tool's live help |\n| Call an installed user tool from an agent | [pm_user_tool](/valdr/docs/valdr-mcp/user-tools/) | Search, describe the selected revision, then run with a stable request ID |\n\nCapabilities give agents reusable instructions. Workflows coordinate steps, waits, and approvals. User tools let agents and workflows run your scripts. MCP lets agents discover and use Valdr’s tools. Combining them lets you reuse a reviewed process without re-explaining each handoff.\n\n## Where to start\n\n- **Comparing capabilities?** Browse [Features](/valdr/docs/features/) for product-level guides before jumping into UI or MCP details.\n- **Building durable processes?** Start with [Valdr Workflows](/valdr/docs/workflows/) for concepts, included workflows, authoring, and recovery.\n- **New to Valdr?** Continue with [Getting Started](/valdr/docs/getting-started/) to install and configure Valdr.\n- **Exploring features?** Browse the [UI guide](/valdr/docs/ui/) to see what each screen does.\n- **Scaling agents?** Learn how [Agent Packs](/valdr/docs/valdr-packs/) let you ship governed workflows without starting from scratch.\n- **Switching models?** [Providers \u0026 Presets](/valdr/docs/valdr-provider-packs/) keep permissions and safety consistent across providers.","description":"Learn how to install, configure, and use Valdr — the governance layer for AI agents.","title":"Documentation"}
Valdr gives you structured, scored, and auditable control over AI agent workflows. Whether you’re setting up for the first time, exploring the UI, or building reusable agent packs — start here.
Product guides for Workspace Knowledge, agent sessions, reviews, and more
Build and operate durable, reviewable agent processes
Install, configure, and start using Valdr
Navigate the web console: dashboard, plans, tasks, agents
The complete MCP tool surface your agents can call
Bundle charters, prompts, and capabilities for reuse
JSON configuration for AI providers and agent provider presets
Fix install, license, MCP, provider, session, and pack issues
Build with Valdr
Choose the interface for the work you want to automate:
| I want to… | Start here | Check before running |
|---|---|---|
| Build a process with steps, waits, and approvals | Definition Authoring | Validate the definition, inspect warnings, then test with safe inputs |
| Turn a script or CLI into a reusable action | Build Your First Tool | Validate the manifest and JSON contract, package it, and review import preflight |
| Let an agent operate my local workspace | Valdr MCP | Connect to the same data directory, discover tools, and read each tool’s live help |
| Call an installed user tool from an agent | pm_user_tool | Search, describe the selected revision, then run with a stable request ID |
Capabilities give agents reusable instructions. Workflows coordinate steps, waits, and approvals. User tools let agents and workflows run your scripts. MCP lets agents discover and use Valdr’s tools. Combining them lets you reuse a reviewed process without re-explaining each handoff.
Where to start
- Comparing capabilities? Browse Features for product-level guides before jumping into UI or MCP details.
- Building durable processes? Start with Valdr Workflows for concepts, included workflows, authoring, and recovery.
- New to Valdr? Continue with Getting Started to install and configure Valdr.
- Exploring features? Browse the UI guide to see what each screen does.
- Scaling agents? Learn how Agent Packs let you ship governed workflows without starting from scratch.
- Switching models? Providers & Presets keep permissions and safety consistent across providers.