Agents break down goals, plan steps, and execute. Self-correct on failure.
Function calling, code execution, web browsing, file system, your custom tools.
Parallel execution, conditional logic, error recovery. Complex agentic flows.
Define an agent in 20 lines. Deploy to the network. Scale automatically.
Short-term conversation, long-term vector memory, structured state.
Trace every step, tool call, and decision. Debug visually.
Token-efficient, with volume discounts and burn incentives.
Plus token cost
Plus underlying token cost
Sandboxed, isolated
Long-term persistence
Pricing is illustrative. Final rates are governed by on-chain parameters and may vary based on network state.
Tier-based access. Higher stakes unlock better economics and more capacity.
| Tier | Required stake | Access |
|---|---|---|
| Builder | 50M QUBIC | Agent SDK, basic workflows |
| Startup | 150M QUBIC | Persistent memory, code exec |
| Business | 500M QUBIC | Custom tools, A/B testing |
| Enterprise | Custom | On-prem, dedicated capacity |
Drop-in compatible with the OpenAI SDK.
from aigarth import Agent, tool
@tool
def search_docs(query: str) -> list[dict]:
return aigarth_client.search(query, limit=5)
agent = Agent(
model="aigarth-reason-1",
tools=[search_docs],
system="You are a research assistant. Be thorough.",
)
result = agent.run("What's the latest on useful proof of staking?")
print(result.answer)Everything in the standard tier, plus the things enterprises need.