API v1 · Developer documentation

Agent Trading Feed API

Documentation for reading Art Argentum LLM Trading and Algorithmic Trading agents through owner-scoped API endpoints.

Read your trading agents from external software

Agent Trading Feed is owner-scoped: an API key can list and read only trading agents belonging to the same Art Argentum account. The current API exposes a unified catalog and separate LLM/Algo detail routes.

The endpoints documented as CURRENT below are mounted now. Narrow live endpoints such as /positions/ or /trades/ are intentionally not presented as active until they are mounted.

From catalog to one agent

  1. Call the owner-scoped agent catalog.
  2. Read the returned engine and agent_slug.
  3. For engine=llm, call the LLM route.
  4. For engine=algo, call the Algo route.

Endpoint reference

GET/api/v1/client/agents/CURRENT
GET/api/v1/client/agents/llm/<agent_slug>/CURRENT
GET/api/v1/client/agents/algo/<agent_slug>/CURRENT

List agents

cURL

curl \
  -H "Authorization: Bearer aa_live_..." \
  https://art-argentum.com/api/v1/client/agents/

PowerShell

$key = "aa_live_..."
$headers = @{ Authorization = "Bearer $key" }

$agents = Invoke-RestMethod `
  -Uri "https://art-argentum.com/api/v1/client/agents/" `
  -Headers $headers

$agents.items | Format-List *

Python

import requests

r = requests.get(
    "https://art-argentum.com/api/v1/client/agents/",
    headers={"Authorization": "Bearer aa_live_..."},
    timeout=30,
)
r.raise_for_status()

for agent in r.json().get("items", []):
    print(agent)

JavaScript

const r = await fetch(
  "https://art-argentum.com/api/v1/client/agents/",
  {
    headers: {
      Authorization: "Bearer aa_live_..."
    }
  }
);

const data = await r.json();
console.log(data.items);

Agent catalog fields

enginellm or algo; selects the detail endpoint.
agent_slugStable API identifier for the agent. Use this exact value in the next request.
display_nameHuman-readable agent name.
statusCurrent project/agent state.
timeframeTrading timeframe when exposed by the agent export.
modelLLM model metadata when applicable.
Example catalog responseDocumentation sample
{
  "ok": true,
  "items": [
    {
      "engine": "llm",
      "agent_slug": "fx-momentum",
      "display_name": "FX Momentum",
      "status": "active",
      "model": "gpt-5.6-sol",
      "timeframe": "H1"
    },
    {
      "engine": "algo",
      "agent_slug": "pair-reversion",
      "display_name": "Pair Reversion",
      "status": "active",
      "timeframe": "M15"
    }
  ]
}

Use the engine returned by the catalog

# Algorithmic Trading
GET /api/v1/client/agents/algo/<agent_slug>/

# LLM Trading
GET /api/v1/client/agents/llm/<agent_slug>/
The current detail routes are broad compatibility exports. Integrations should consume only fields they actually need rather than assuming every internal project field is a permanent external contract.

Authentication, billing and errors

Send Authorization: Bearer aa_live_... on every request. For YouTube source files, successful concrete-file access costs $0.01; discovery is free.

Common responses: 200 success · 401 invalid key · 402 insufficient balance · 404 missing or unauthorized resource · 405 method not allowed.

See API overview, billing headers and common response semantics →