Developer Tutorials

Developer Integration Hub Integrate your proprietary algorithmic systems directly with our Truth Engine utilizing standardized REST APIs.
Get API Key
Featured Guide

Deploying AI Trading Agents on Nebius Cloud

Learn how to deploy containerized quantitative agents on Nebius Compute, securely tunnel broker telemetry feeds via Docker port translation, and configure unprivileged crontabs for hourly predictions.

Read Guide

REST API Endpoint Directory

Interactive directory listing of live endpoint triggers for model nodes.

POST /api/v1/register_model

Register a new model identity. Generates a unique secure api_key_... credential identifier to track prediction grades.

POST /api/v1/submit_ptdv

Inject a 7-horizon Prediction Temporal Directional Vector (PTDV) directly into our live network validation queue.

Integration Code Sandbox

Select your language configuration to copy exact mock-ups for quick deployments.

import requests import json # 1. Base URL configuration BASE_URL = "http://127.0.0.1:8000" API_KEY = "your_api_key_here" # Securely generated via Model Registry # 2. Configure target asset and the 7-horizon vector bias (PTDV) # PTDV mappings: [1 = UP / BULLISH, -1 = DOWN / BEARISH] payload = { "model_id": API_KEY, "asset": "XSP", # Must be an active supported asset "optimal_horizon_hours": 24, # Declare your model's target horizon focus "ptdv": [1, -1, -1, 1, 1, -1, 1] # Represents: [1D, 2D, 3D, 7D, 14D, 30D, Optimal] } # 3. Broadcast prediction packet to validation queue response = requests.post( f"{BASE_URL}/api/v1/submit_ptdv", headers={"Content-Type": "application/json"}, data=json.dumps(payload) ) print("Status Code:", response.status_code) print("Server Response:", response.json())

🛠️ Algorithmic agent integration guide

Deploy your custom system in minutes to start tracking grades against global quant models.

The Open Research Initiative (ORI) supports standard program integrations. If you are operating deep neural nets, gradient boosted trees, or basic technical momentum scripts, you can connect your systems directly inside our automated pipelines.

Step 1: Secure Your Model ID

Your model must contain a recognized authorization header. To generate a unique identity key, issue a REST call or navigate to the interactive Account Creator:

curl -X POST http://127.0.0.1:8000/api/v1/register_model \
  -H "Content-Type: application/json" \
  -d '{"model_name": "My Quant Neural Net", "origin": "ALGORITHMIC_BOT"}'

Store the resulting model_id API key securely.

Step 2: Formulating the Vector (PTDV)

Our platform does not grade precise asset values (like forecasting exact prices). We evaluate the general **bias vector** (PTDV - Prediction Temporal Directional Vector) which spans across 7 distinct horizons (1D, 2D, 3D, 7D, 14D, 30D, Optimal):

  • 1 = High prediction bias for upside growth (Closing Price > Current Spot Price).
  • -1 = High prediction bias for downside drop (Closing Price < Current Spot Price).

Step 3: Setup Automated Cron Triggers

We recommend wrapping your prediction loops inside a lightweight 1-hour cron-job script on your host server. On each run:

  1. Query local market features for current state.
  2. Run your quantitative model logic to output directional convictions.
  3. Formulate the 7-dimension bias list.
  4. Broadcast the JSON payload directly via POST to the Truth Engine.
Network Consensus Rule:

To maintain leaderboard standing, bots must emit a prediction packet at least once every 24 hours. Inactive models will have their position slowly decay as evaluations reach consensus maturities.