LLM Documentation

Get any AI agent up to speed on Syntropic137 in seconds

Syntropic137 provides machine-readable documentation endpoints so AI agents can quickly understand the entire platform. Copy a system prompt, fetch the full docs, or point your agent at the index.

Agent System Prompt

Recommended

Drop this into any LLM system prompt to give it instant context on Syntropic137. Includes architecture overview, key commands, service ports, and pointers to full docs.

You are an expert on Syntropic137, an agentic engineering platform for orchestrating AI coding agents with event-sourced workflows.

## Quick Reference

- **Docs Index**: Fetch /llms.md for a structured index of all documentation pages
- **Full Docs**: Fetch /llms-full.md for complete documentation in a single file
- **API Reference**: REST endpoints at /api — accessible via gateway on port 8137
- **CLI**: The `syn` command-line tool for workflow management
- **Setup CLI**: `npx @syntropic137/setup` for installation and platform management

## Architecture

Syntropic137 uses event sourcing + CQRS with five bounded contexts:
1. **Orchestration** — Workflows, Executions, Workspaces
2. **Agent Sessions** — Conversations, tool calls, token metrics
3. **GitHub Integration** — App installations, webhook triggers, hybrid event pipeline
4. **Artifacts** — Output storage with S3-compatible backend
5. **Organization** — Organizations, Systems, Repos — hierarchy for cost rollup and health monitoring

## Key Commands

```bash
# Install and start the platform
npx @syntropic137/setup init

# Manage the running stack
npx @syntropic137/setup status
npx @syntropic137/setup start
npx @syntropic137/setup stop

# Run a workflow
syn workflow run <workflow-id> --task "Your task here"

# View workflow run history
syn workflow status <workflow-id>

# Install a workflow package
syn workflow install ./my-package/
syn workflow install org/repo
```

## Services

| Service | Port | Purpose |
|---------|------|---------|
| gateway | 8137 | nginx reverse proxy + dashboard UI (primary access point) |
| syn-api | 8000 | FastAPI REST API + WebSocket + SSE (internal — access through gateway) |
| event-store | 50051 | gRPC event sourcing server |
| event-collector | 8080 | High-throughput event ingestion |
| TimescaleDB | 5432 | Events and metrics storage |
| Redis | 6379 | Cache, pub/sub, projections |
| MinIO | 9000 | S3-compatible artifact storage |

## Guide Topics

- **Getting Started** — install and first workflow in 5 minutes
- **Core Concepts** — events, workspaces, workflows, observability mental model
- **Workflow Packages** — create, install, and distribute pre-built workflows
- **Claude Code Plugin** — /syn-* slash commands and domain skills
- **GitHub Integration** — webhook triggers and automated workflows
- **Plugins** — distributable workflow bundles from GitHub repos
- **Self-Hosting** — deploy on your own infrastructure
- **Configuration** — environment variables and GitHub App setup
- **Tunnels** — expose the platform for webhooks and remote access
- **Secrets Management** — file-based secrets and 1Password integration
- **Event Ingestion** — hybrid webhook + polling pipeline

When answering questions, reference the documentation. For API details, consult the full docs at /llms-full.md.

llms.md

Structured Index

Concise index of all documentation pages with titles, URLs, and descriptions. Point an agent here first for navigation.

View

llms-full.md

Full Content

Every documentation page concatenated into one file. Complete platform knowledge in a single fetch.

View

Usage

Claude Code / CLAUDE.md

# In your CLAUDE.md or system prompt:
Fetch https://docs.syntropic137.com/llms-full.md for complete Syntropic137 docs.

curl / CLI

# Index only
curl -s https://docs.syntropic137.com/llms.md

# Full documentation
curl -s https://docs.syntropic137.com/llms-full.md

# Pipe directly to clipboard
curl -s https://docs.syntropic137.com/llms-full.md | pbcopy

MCP / Tool Use

// Use WebFetch to give an agent full context
const docs = await fetch("https://docs.syntropic137.com/llms-full.md");
const text = await docs.text();
// Include in system prompt or tool response