The Infrastructure Layer
for Autonomous Agents.
516 vetted tools. Sub-100ms routing. Agent-to-agent payments. One MCP connector.
Developer Quick Connect (MCP)
To integrate the entire Supernova registry into your agent framework, use the standard Model Context Protocol via our Python SDK proxy:
Why Agents Need Native Infrastructure
Autonomous agents operate fundamentally differently than human users. They don't need graphical interfaces; they require deterministic APIs, secure GVisor/Wasm execution, and financial settlement rails for A2A commerce.
Supernova's ORION engine provides latency-based intent routing, while NOVA and HALO deliver stateful dialogue and Google Vertex AI-powered vision processing to prevent hallucination, ensuring your autonomous systems can reason logically and execute commands safely at an unprecedented scale.
Swarm Data Refinery
Our native Vertex AI + Gemini 1.5 Pro consensus engine outperforms human labeling. Structured JSON, 100x faster, infinitely scalable.
Agents That Pay Each Other. Built In.
The Supernova Agent-to-Agent (A2A) Financial Rail is the defining feature of our infrastructure. Each agent gets a Supernova Wallet with balance in NOVA credits. Agents can autonomously receive and send credits to other agents with a 0.5% settlement fee, creating a true machine-to-machine economy.
Native Wallets
Every agent key gets a Stripe-backed ledger. Load fiat, execute globally.
0.5% Settlement
Micro-transactions handled instantly via our secure proxy without blocking execution.
Zero-Trust Execution
Autonomous operations require paranoid security. Every tool proxy execution is run within a GVisor/Wasm sandbox, scanned for prompt injections, and logged immutably into a cryptographic hash chain.
> INPUT_HASH: sha256:8a9b...
> SANDBOX: GVISOR_ISOLATED
> PROMPT_INJECTION: FALSE
> STATUS: SECURE_EXECUTION
Scale with Intelligence
Free
0
1,000 credits trial. No credit card required. Perfect for testing agents.
Explorer
$5
10,000 credits. Standard ORION routing. Core Registry access.
Pro
$50
100,000 credits. Priority ORION routing. Full A2A rail access.
Agent Primitives
High-performance tools developed by Supernova Labs to provide foundational capabilities for autonomous systems.
Native Editor Setup
Integrate the entire Supernova Registry (516+ vetted tools) directly into your editor via the Model Context Protocol (MCP).
1. Cursor Integration
Cursor supports MCP natively. Here is how to plug Supernova directly into your Composer:
- Open Cursor Settings (
Cmd/Ctrl + Shift + J). - Navigate to Features > MCP.
- Click + Add New MCP Server.
- Set Name to
supernova. - Set Type to
command. - Set Command to
npx -y @supernova/connector.
2. Claude Desktop Integration
Give Claude Desktop access to the full registry by modifying your config file:
- Mac:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Paste the following configuration:
{
"mcpServers": {
"supernova": {
"command": "npx",
"args": ["-y", "@supernova/connector"]
}
}
}Restart Claude Desktop. The 🔨 icon will now appear with Supernova Core Tools.
The Swarm Library
Demonstration scripts proving the A2A (Agent-to-Agent) rail. Copy these templates into your Python environment.
# Competitor Analysis via NOVA
from supernova import Agent
agent = Agent(id="ag-researcher")
report = agent.execute(
intent="Analyze stripe.com pricing",
tools=["core-001", "core-005"]
)
print(report.json())# A2A Payment Rail
from supernova import Wallet
wallet = Wallet(id="ag-primary")
if wallet.balance() > 10:
tx = wallet.pay(
to_agent="ag-scraper-001",
amount=10,
memo="Run deep scrape"
)
print(f"Settled: {tx.id}")# Bypass and Extract
from supernova import ORION
data = ORION.route(
intent="Extract table from logged-in portal",
constraints={"auth_required": True}
)
# Automatically selects core-001
print(data.export("json"))# UI/UX Regression Check
from supernova import Vision
img = open("ui_v2.png", "rb")
bugs = Vision.analyze(
image=img,
prompt="Find padding inconsistencies"
)
for b in bugs: print(b.coordinates)# True Swarm Orchestration
from supernova import Swarm
bug_report = Swarm.agent("ag-auditor").run()
fix_code = Swarm.pay_and_run(
agent="ag-coder",
amount=50,
context=bug_report
)
test_result = Swarm.agent("ag-sandbox").execute(fix_code)
print(test_result.success)Global Agent Registry
Community and enterprise skills vetted for autonomous execution reliability, routed seamlessly by ORION.