Skip to content
lacausecrypto edited this page Apr 6, 2026 · 10 revisions

OCC — Claude Chain Orchestrator Wiki

Welcome to the OCC documentation. OCC is an open-source multi-model workflow orchestration engine — define multi-step chains in YAML, execute them with parallel processing, and access everything via MCP, REST API, or the React dashboard.

49 test files, 1789 tests, 102 REST endpoints, 28 MCP tools, 6 LLM providers.

Quick Navigation

Getting Started

  • Getting Started — Install, configure, run your first chain in 5 minutes
  • Configuration — Environment variables, .mcp.json, Docker, deployment

Chain Authoring

  • Chain Format — Complete YAML reference for chain definitions
  • Step Types — All 11 step types explained with examples
  • Pre-Tools — 29 pre-tool types (web search, API calls, MCP, bash, state, vector, etc.)
  • Pipelines — Chain multiple chains together

Tools & API

  • CLI — Command-line interface: run, validate, dry-run, cost estimate (17 commands)
  • REST API — 102 HTTP endpoints with request/response examples
  • MCP Tools — 28 MCP tools for Claude Code / Claude Desktop
  • MCP Client — Consume 10,000+ external MCP servers in chains

Advanced

  • Architecture — 18 modules: executor, queue, storage, linter, MCP client, providers, etc.
  • Token Optimization — Minimize cost with caching, transforms, context strategies
  • Cookbook — Practical recipes and patterns

Help

  • FAQ — Common questions and troubleshooting

Key Concepts

Concept Description
Chain A YAML file defining a multi-step workflow. Each step is an LLM call, data transform, or control flow.
Step One unit of work in a chain. 11 types: agent, router, evaluator, gate, transform, loop, merge, browser, subchain, debate, webhook.
Pipeline Multiple chains executed in sequence/parallel, with outputs passed between them.
Pre-tool Data injection before a step — 29 types including web search, HTTP fetch, file read, bash, env var, MCP call, vector query, state management, semantic cache. Supports on_error (inject/skip/fail), retry, parallel execution.
Variable Outputs stored as output_var and interpolated into prompts via {variable_name}. Supports {key.subkey} and {key|"fallback"}.
Wave A group of steps with no dependencies between them — they execute in parallel.
Gate A step that pauses execution until a human approves or rejects via the API or dashboard.
Queue Persistent SQLite job queue — if all workers are busy, new executions are queued with priority.
Checkpoint Each step result is saved to SQLite immediately. Crash recovery resumes from last completed step.
Execution A running instance of a chain, with its own ID, status, step results, and SSE stream.
Provider An LLM backend: Claude CLI, OpenRouter, OpenAI, Ollama (local), HuggingFace, or any OpenAI-compatible API.
BLOB The visual canvas — an organic workflow graph builder with conversational AI.
Workflow Chat A chat-based chain builder with agentic actions (run, stop, debug, modify steps).

Project Stats

Metric Value
REST endpoints 102
MCP tools 28
Pre-tool types 29
Step types 11
LLM providers 6 (Claude, OpenRouter, OpenAI, Ollama, HuggingFace, Custom)
Example chains 16
Example pipelines 5
Test files 49
Tests 1789
CI platforms Ubuntu, macOS, Windows × Node 20 + 22

Clone this wiki locally