Agents API - Shannon

Overview

The Agents API provides direct execution of single-purpose agents (also called “Quick Tools”) without LLM orchestration. Each agent wraps a specific tool and returns structured results. Key Differences from Task API:

Agents API: Direct tool execution, no AI orchestration, synchronous results
Task API: Multi-step workflows, LLM planning, asynchronous execution

All agent executions are asynchronous (return task_id), even though agents perform single operations. Use the Task Status API to retrieve results.

Base URL

http://localhost:8080/api/v1/agents

Production: https://api-dev.shannon.run/api/v1/agents

Authentication

Required: Yes Include API key in header:

X-API-Key: sk_your_api_key

Development Default: Authentication is disabled when GATEWAY_SKIP_AUTH=1 is set.

Endpoints

List Agents

GET /api/v1/agents Returns all available agents with their schemas and metadata.

Request

curl http://localhost:8080/api/v1/agents \
  -H "X-API-Key: sk_test_123456"

Response

{
  "agents": [
    {
      "id": "serp-ads",
      "name": "SERP Ads Extract",
      "description": "Extract paid ads from Google Search results for given keywords",
      "category": "ads_research",
      "tool": "ads_serp_extract",
      "input_schema": {
        "required": ["keywords"],
        "properties": {
          "keywords": {
            "type": "string",
            "description": "Search keywords to find ads for"
          },
          "country": {
            "type": "string",
            "default": "us",
            "description": "Country code (us, jp, uk, etc.)"
          }
        }
      },
      "cost_per_call": 0.015
    }
  ],
  "count": 16
}

Response Fields

Field	Type	Description
`agents`	array	List of agent definitions
`count`	integer	Total number of agents

Agent Object:

Field	Type	Description
`id`	string	Agent identifier (used in execute endpoint)
`name`	string	Human-readable name
`description`	string	What the agent does
`category`	string	Agent category (ads_research, financial, etc.)
`tool`	string	Underlying tool name
`input_schema`	object	JSON schema for input validation
`cost_per_call`	number	Estimated cost per execution (USD)

Get Agent Details

GET /api/v1/agents/ Returns details for a specific agent, including its input schema.

Request

curl http://localhost:8080/api/v1/agents/serp-ads \
  -H "X-API-Key: sk_test_123456"

Response

{
  "id": "serp-ads",
  "name": "SERP Ads Extract",
  "description": "Extract paid ads from Google Search results for given keywords",
  "category": "ads_research",
  "tool": "ads_serp_extract",
  "input_schema": {
    "required": ["keywords"],
    "properties": {
      "keywords": {
        "type": "string",
        "description": "Search keywords to find ads for"
      },
      "country": {
        "type": "string",
        "default": "us"
      },
      "device": {
        "type": "string",
        "enum": ["desktop", "mobile", "tablet"],
        "default": "desktop"
      }
    }
  },
  "cost_per_call": 0.015
}

Error Responses

404 Not Found - Agent does not exist:

{
  "error": "Agent not found: invalid-agent-id"
}

Execute Agent

POST /api/v1/agents/ Executes a specific agent with provided input. Returns a task ID immediately; the agent runs asynchronously.

Request Headers

Header	Required	Description
`X-API-Key`	Yes	API authentication key
`Content-Type`	Yes	Must be `application/json`

Request Body

Parameter	Type	Required	Description
`input`	object	Yes	Agent-specific input parameters (validated against `input_schema`)
`session_id`	string	No	Session identifier for tracking (auto-generated if omitted)
`stream`	boolean	No	Enable streaming (reserved for future use)

Example: Execute SERP Ads Agent

curl -X POST http://localhost:8080/api/v1/agents/serp-ads \
  -H "X-API-Key: sk_test_123456" \
  -H "Content-Type: application/json" \
  -d '{
    "session_id": "user-session-123",
    "input": {
      "keywords": "best credit cards",
      "country": "us",
      "device": "desktop"
    }
  }'

Response

Status: 202 Accepted Headers:

X-Workflow-ID: Temporal workflow identifier
X-Session-ID: Session identifier

Body:

{
  "task_id": "task-abc123",
  "agent_id": "serp-ads",
  "status": "STATUS_CODE_OK",
  "created_at": "2026-02-15T10:30:00Z"
}

Retrieve Results

Use the Get Task Status endpoint with the returned task_id:

curl http://localhost:8080/api/v1/tasks/task-abc123 \
  -H "X-API-Key: sk_test_123456"

Response when completed:

{
  "task_id": "task-abc123",
  "status": "TASK_STATUS_COMPLETED",
  "result": "{\"keywords\":\"best credit cards\",\"total_ads\":5,\"ads\":[...]}",
  "response": {
    "keywords": "best credit cards",
    "total_ads": 5,
    "ads": [
      {
        "position": 1,
        "title": "Best Credit Cards 2026",
        "description": "Compare top credit cards...",
        "link": "https://example.com/cards",
        "domain": "example.com"
      }
    ],
    "cost_usd": 0.015,
    "timestamp": "2026-02-15T10:30:00Z"
  },
  "model_used": "claude-sonnet-4-20250514",
  "provider": "anthropic",
  "usage": {
    "total_tokens": 1250,
    "input_tokens": 800,
    "output_tokens": 450,
    "estimated_cost": 0.018
  }
}

Error Responses

400 Bad Request - Invalid input:

{
  "error": "input validation failed: missing required field: keywords"
}

404 Not Found - Agent does not exist:

{
  "error": "Agent not found: invalid-agent"
}

429 Too Many Requests - Rate limit exceeded:

{
  "error": "Rate limit exceeded"
}

Input Validation

All agent inputs are validated against the agent’s input_schema before execution. Validation Rules:

Required fields must be present and non-null
Type checking - strings, integers, booleans, arrays, objects
Enum validation - values must be in allowed list
Unknown fields - rejected for security (not in schema)

Example Schema:

{
  "required": ["keywords"],
  "properties": {
    "keywords": {
      "type": "string",
      "description": "Search keywords"
    },
    "device": {
      "type": "string",
      "enum": ["desktop", "mobile", "tablet"],
      "default": "desktop"
    },
    "max_results": {
      "type": "integer",
      "default": 10
    }
  }
}

Valid input:

{
  "keywords": "best shoes",
  "device": "mobile",
  "max_results": 5
}

Invalid input (missing required field):

{
  "device": "mobile"
}

Error: "input validation failed: missing required field: keywords" Invalid input (unknown field):

{
  "keywords": "shoes",
  "unknown_field": "value"
}

Error: "input validation failed: unknown field: unknown_field (not defined in agent schema)"

Available Agents

Shannon provides 16+ specialized agents across multiple categories. For a complete catalog of available agents with detailed schemas and examples, see:

Ads Research Agents

10 agents for competitive advertising analysis

Financial Research Agents

4 agents for stock news and sentiment analysis

Quick Reference by Category

Ads Research (10 agents):

serp-ads - Extract Google paid ads
yahoo-jp-ads - Extract Yahoo Japan sponsored ads
meta-ad-library - Search Meta Ad Library (Facebook/Instagram)
competitor-discover - Find competitor advertisers
ads-transparency - Multi-platform ad transparency data
lp-visual-analyze - Screenshot and analyze landing pages
lp-batch-analyze - Batch analyze multiple landing pages
ad-creative-analyze - Analyze ad copy patterns
keyword-extract - Extract search keywords from text
browser-screenshot - Capture webpage screenshots

Financial Tools (4 agents):

sec-filings - SEC EDGAR filings lookup
twitter-sentiment - X/Twitter sentiment via xAI
alpaca-news - Stock news from Alpaca Markets
news-aggregator - Multi-source news aggregation

Unified Task API Alternative

You can also execute agents through the unified Task API using the context.agent parameter:

curl -X POST http://localhost:8080/api/v1/tasks \
  -H "X-API-Key: sk_test_123456" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "NVDA",
    "session_id": "user-session-123",
    "context": {
      "agent": "sec-filings",
      "agent_input": {
        "ticker": "NVDA",
        "days_back": 90
      }
    }
  }'

Both approaches are equivalent:

Dedicated endpoint: POST /api/v1/agents/{id}
Unified endpoint: POST /api/v1/tasks with context.agent

Best Practices

1. Validate Inputs Before Submission

Use the GET endpoint to retrieve the agent’s schema, then validate your input client-side:

import httpx

# Get schema
agent = httpx.get(
    "http://localhost:8080/api/v1/agents/serp-ads",
    headers={"X-API-Key": "sk_test_123456"}
).json()

# Validate required fields
required = agent["input_schema"]["required"]
input_data = {"keywords": "shoes", "device": "mobile"}

for field in required:
    if field not in input_data:
        raise ValueError(f"Missing required field: {field}")

2. Handle Async Results

All agents return task IDs immediately. Poll for results:

import time

# Execute agent
response = httpx.post(
    "http://localhost:8080/api/v1/agents/serp-ads",
    headers={"X-API-Key": "sk_test_123456"},
    json={"input": {"keywords": "shoes"}}
).json()

task_id = response["task_id"]

# Poll for completion
while True:
    status = httpx.get(
        f"http://localhost:8080/api/v1/tasks/{task_id}",
        headers={"X-API-Key": "sk_test_123456"}
    ).json()

    if status["status"] == "TASK_STATUS_COMPLETED":
        print(status["response"])
        break

    time.sleep(2)

3. Use Sessions for Context

Reuse session_id across related agent calls:

session_id = "user-123-analysis"

# First agent call
httpx.post(..., json={
    "session_id": session_id,
    "input": {"keywords": "shoes"}
})

# Second agent call (same session)
httpx.post(..., json={
    "session_id": session_id,
    "input": {"urls": ["https://example.com"]}
})

4. Check Cost Estimates

Before executing expensive agents, check cost_per_call:

agent = httpx.get(
    "http://localhost:8080/api/v1/agents/twitter-sentiment",
    headers={"X-API-Key": "sk_test_123456"}
).json()

if agent["cost_per_call"] > 0.10:
    print(f"Warning: High cost agent (${agent['cost_per_call']})")

Code Examples

Python with httpx

import httpx

# List all agents
agents = httpx.get(
    "http://localhost:8080/api/v1/agents",
    headers={"X-API-Key": "sk_test_123456"}
).json()

print(f"Found {agents['count']} agents")

# Execute specific agent
response = httpx.post(
    "http://localhost:8080/api/v1/agents/serp-ads",
    headers={
        "X-API-Key": "sk_test_123456",
        "Content-Type": "application/json"
    },
    json={
        "session_id": "my-session",
        "input": {
            "keywords": "running shoes",
            "country": "us",
            "device": "mobile"
        }
    }
).json()

print(f"Task ID: {response['task_id']}")

JavaScript/Node.js

const axios = require('axios');

// Get agent details
const agent = await axios.get(
  'http://localhost:8080/api/v1/agents/serp-ads',
  {
    headers: { 'X-API-Key': 'sk_test_123456' }
  }
);

console.log('Agent schema:', agent.data.input_schema);

// Execute agent
const response = await axios.post(
  'http://localhost:8080/api/v1/agents/serp-ads',
  {
    session_id: 'my-session',
    input: {
      keywords: 'running shoes',
      country: 'us'
    }
  },
  {
    headers: {
      'X-API-Key': 'sk_test_123456',
      'Content-Type': 'application/json'
    }
  }
);

console.log('Task ID:', response.data.task_id);

Go

package main

import (
    "bytes"
    "encoding/json"
    "fmt"
    "net/http"
)

type AgentExecuteRequest struct {
    SessionID string                 `json:"session_id,omitempty"`
    Input     map[string]interface{} `json:"input"`
}

func executeAgent(agentID string, input map[string]interface{}) (string, error) {
    req := AgentExecuteRequest{
        SessionID: "my-session",
        Input:     input,
    }

    body, _ := json.Marshal(req)

    httpReq, _ := http.NewRequest(
        "POST",
        fmt.Sprintf("http://localhost:8080/api/v1/agents/%s", agentID),
        bytes.NewBuffer(body),
    )

    httpReq.Header.Set("X-API-Key", "sk_test_123456")
    httpReq.Header.Set("Content-Type", "application/json")

    client := &http.Client{}
    resp, err := client.Do(httpReq)
    if err != nil {
        return "", err
    }
    defer resp.Body.Close()

    var result map[string]interface{}
    json.NewDecoder(resp.Body).Decode(&result)

    return result["task_id"].(string), nil
}

func main() {
    taskID, _ := executeAgent("serp-ads", map[string]interface{}{
        "keywords": "running shoes",
        "country":  "us",
    })

    fmt.Printf("Task ID: %s\n", taskID)
}

Submit Task

Unified task submission with AI orchestration

Get Task Status

Retrieve agent execution results

Ads Research Agents

Complete ads research agent catalog

Financial Agents

Financial research agent catalog

Getting Started

Tasks

Streaming

Agents

Sessions

Swarm

Skills

Schedules

HITL Review

Reference

​Overview

​Base URL

​Authentication

​Endpoints

​List Agents

​Request

​Response

​Response Fields

​Get Agent Details

​Request

​Response

​Error Responses

​Execute Agent

​Request Headers

​Request Body

​Example: Execute SERP Ads Agent

​Response

​Retrieve Results

​Error Responses

​Input Validation

​Available Agents

Ads Research Agents

Financial Research Agents

​Quick Reference by Category

​Unified Task API Alternative

​Best Practices

​1. Validate Inputs Before Submission

​2. Handle Async Results

​3. Use Sessions for Context

​4. Check Cost Estimates

​Code Examples

​Python with httpx

​JavaScript/Node.js

​Go

​Related Endpoints

Submit Task

Get Task Status

Ads Research Agents

Financial Agents

Overview

Base URL

Authentication

Endpoints

List Agents

Request

Response

Response Fields

Get Agent Details

Request

Response

Error Responses

Execute Agent

Request Headers

Request Body

Example: Execute SERP Ads Agent

Response

Retrieve Results

Error Responses

Input Validation

Available Agents

Quick Reference by Category

Unified Task API Alternative

Best Practices

1. Validate Inputs Before Submission

2. Handle Async Results

3. Use Sessions for Context

4. Check Cost Estimates

Code Examples

Python with httpx

JavaScript/Node.js

Go

Related Endpoints