Chains API

Agent chains let you run multiple agents in sequence, passing output from one to the next.

List Chain Presets

Get available pre-built chain templates.

GET /chains
Authorization: Bearer <token>

Response (200):

{
  "chains": [
    {
      "id": "code-review-pipeline",
      "name": "Code Review Pipeline",
      "description": "Full code review with testing and documentation",
      "steps": ["code-review", "testing-validation", "documentation"],
      "tier": "free"
    },
    {
      "id": "security-audit",
      "name": "Security Audit",
      "description": "Comprehensive security analysis",
      "steps": ["code-review", "365-security", "deployment"],
      "tier": "pro"
    }
  ]
}

---

Execute Chain

Run a chain of agents sequentially.

POST /chains/execute
Authorization: Bearer <token>
Content-Type: application/json

Request Body:

{
  "steps": [
    {"agent": "code-review"},
    {"agent": "testing-validation"},
    {"agent": "documentation"}
  ],
  "context": "function fetchUsers() { ... }",
  "provider": "deepseek"
}

Field	Type	Required	Description
steps	array	✅	Array of {agent} objects defining the chain
context	string	✅	Initial context passed to the first agent
provider	string	❌	LLM provider (default: deepseek)
chain	string	❌	Use a preset chain ID instead of manual steps

Response (200 SSE stream):

data: {"type":"chain_start","steps":3}

data: {"type":"step_start","step":1,"agent":"code-review"}
data: {"type":"chunk","content":"## Code Review Results..."}
data: {"type":"step_done","step":1,"agent":"code-review","duration":"1.4s"}

data: {"type":"step_start","step":2,"agent":"testing-validation"}
data: {"type":"chunk","content":"## Generated Tests..."}
data: {"type":"step_done","step":2,"agent":"testing-validation","duration":"2.1s"}

data: {"type":"step_start","step":3,"agent":"documentation"}
data: {"type":"chunk","content":"## API Documentation..."}
data: {"type":"step_done","step":3,"agent":"documentation","duration":"1.8s"}

data: {"type":"chain_done","totalDuration":"5.3s","totalTokens":2847}

---

Example: Full Chain Execution

TOKEN="your-jwt-token"

# Run a 3-step chain
curl -N -X POST https://api.lunaos.ai/chains/execute \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "steps": [
      {"agent": "requirements-analyzer"},
      {"agent": "design-architect"},
      {"agent": "code-review"}
    ],
    "context": "Build a real-time collaborative document editor with conflict resolution"
  }'

Chain Rules

• Max steps: 5 agents per chain
• Output passing: Each step's output is appended to the context for the next step
• Pro agents in chains: If any step uses a Pro agent, the user must be on the Pro plan
• Execution count: Each chain counts as one execution toward monthly limits