> ## Documentation Index > Fetch the complete documentation index at: https://docs.orisu.ai/llms.txt > Use this file to discover all available pages before exploring further. # Resources > Stable orisu:// URIs the assistant reads to discover state: agents, runs, nodes, models, and the workflow skill itself. MCP resources are cacheable reads, pure GETs against a `orisu://` URI. They're how the assistant inspects state without spending a tool call. ## Discovery resources | URI | Returns | | ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `orisu://nodes` | Lists every node type Orisu supports (\~56), grouped by category, with ports (including loop/fan-out flags) + descriptions. | | `orisu://nodes/{type}` | Full definition for a single node type: `config.schema_fields` (every key with type, enum values, ranges, descriptions), port shapes, capability kind, compatible models. | | `orisu://models` | The model catalog: every model variant Orisu routes to, with its capabilities and pricing. | | `orisu://models/{variant_id}` | One model variant with full settings, pricing, and `best_for` / `when_to_use` selection guidance. | | `orisu://templates/{id}` | One published template with its actual graph (nodes, edges, scrubbed config) to model a new workflow on. Find ids via `search({ kind: 'template' })`. | Read `orisu://nodes` once at the start of any build session. The assistant needs it to pick node types. ## State resources | URI | Returns | | ------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `orisu://agents` | List of agents in the bound organization. | | `orisu://agents/{id}` | One agent: current graph plus latest version metadata. | | `orisu://agents/{id}/versions` | Version history for an agent. | | `orisu://apps` | List of published apps. | | `orisu://apps/{id}` | One app's metadata + input schema. | | `orisu://runs/{id}` | A run's full snapshot: top-level status, per-step status, and per-step outputs. Used by `wait_for_run` internally; you can poll it manually for progress mid-run. | ## Account resources | URI | Returns | | ------------------------- | --------------------------------------------------------------------------------------------------------------------------------------- | | `orisu://assets` | Recent assets the organization has produced. | | `orisu://brand-kits` | Brand kits available in the organization. | | `orisu://brand-kits/{id}` | One brand kit's full content. | | `orisu://integrations` | Connected external accounts (Google Drive, scrapers, etc.) plus `connect_url` — the exact page to send the user to when one is missing. | | `orisu://credits` | Current credit balance + recent transactions. | ## Reference resources | URI | Returns | | -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | `orisu://skills/workflows` | The [orisu-workflows](/mcp/tools#workflow-skill) agent skill: the canonical playbook for build/run/iterate. Server `instructions` direct new agents here. Always reflects the latest skill version because it's served straight off the MCP server (no separate CDN to drift). | ## UI resources | URI | Returns | | ---------------------------- | ---------------------------------------------------------------------------------------------------- | | `ui://orisu/workflow-canvas` | HTML shim that the chat client's MCP-Apps sandbox loads to render the [in-chat canvas](/mcp/canvas). | ## When to read what The [workflow skill](/mcp/tools#workflow-skill) opens with the "Discovery: always do this before designing" section that maps tasks to the resources you should read first. Short version: * **Building a new workflow** → read `orisu://nodes`, then `orisu://nodes/{type}` for each node type you'll use * **Editing an existing agent** → read `orisu://agents/{id}` (or call `view_agent`) * **Picking a model** → read `orisu://models` or call `suggest_model` * **Checking a run** → call `wait_for_run` (single round-trip) instead of polling `orisu://runs/{id}` * **Loading a brand context** → read `orisu://brand-kits` then `orisu://brand-kits/{id}` ## See also The 20+ tools that mutate or trigger state. What `RESOURCE_NOT_FOUND` and friends mean.