AI Agent Stack for Seoul and South Korea's Startup Capital

Founders around Seoul already know which model they want. They still open a second account for market data, a third for the agent run, and a laptop for the script. SandBase collapses that stack:

• More than 2,000 models and data APIs, across 100+ platforms, against one balance
• Agents that stay on the account and Sessions you start, compatible with OpenAI Agents and OpenAgentCore
• Sandboxes on the E2B interface when the work has to read a file or execute a command
• SKILL.md so Cursor, Claude Code, or Codex can find the catalog

Opening an Agent or a sandbox does not require a charge. Eligible data APIs share a pool of 100 free calls a day per organization. The pool resets at 00:00 UTC. Model traffic is billed apart from that pool.

Send traffic to https://api.sandbase.ai/v1 and authenticate with Authorization: Bearer $SANDBASE_API_KEY.

These public pages describe the city and the country. SandBase is not announcing a Seoul address.

• In the GSER 2026 Top 40, Seoul is #9 globally and #3 in Asia. The Seoul section of that report says the startup ecosystem is entering a new phase of global competitiveness, underpinned by venture capital and R&D. Startup Genome’s Seoul profile is the public page for that ecosystem.
• Seoul National University’s AI Institute describes AIIS as an intercollegiate research institute that integrates and supports AI research, and as a hub for Core AI and X+AI work across disciplines.
• KAIST’s Beyond AI strategy calls for education and research that advances humanistic, democratic, and agentic AI, and for deep-tech innovation that changes the startup environment together with industry. That is a national science and technology strategy, not a SandBase location.

SandBase is the API under that mix of a top-ranked startup city and university AI labs: one key, rather than a new vendor for every model or dataset.

1. Mint one key in the Console for the Korean organization.
2. Register the Agent through POST /v1/agents or the OpenAI SDK, including the model, the instructions, and the tools.
3. Wait until projection_status is ready, then create the Session and hand it the task.
4. Point the E2B SDK at https://sandbox.sandbase.ai if the task needs files or a script.
5. Or give the assistant SKILL.md and let it discover a tool before it runs one.

Longer paths: Docs · Agents API · Sandbox quickstart

The OpenAI Python SDK pattern from the SandBase docs, with instructions for Seoul and South Korean readers.

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["SANDBASE_API_KEY"],
    base_url="https://api.sandbase.ai/v1",
)

agent = client.beta.agents.create(
    model="deepseek/deepseek-v4-flash",
    name="Seoul startup brief",
    instructions=(
        "Write for a Seoul and South Korean startup audience. "
        "Compare more than one source and cite each."
    ),
    metadata={"city": "seoul", "team": "korea"},
)

print(agent.id, agent.version)  # agt_..., 1
# Open a Session only after projection_status is ready.

curl -X POST https://api.sandbase.ai/v1/agents/sessions \
  -H "Authorization: Bearer $SANDBASE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"agent_id": "agt_..."}'

Keep scripts and files in a sandbox. Start with the Sandbox quickstart.

If you searched for an AI agent API Seoul teams in South Korea can adopt while keeping the OpenAI Agents SDK, SandBase is that API.

Use these public pages. Do not add a SandBase office, a phone number, or a latency figure for Seoul:

• GSER 2026
• Startup Genome — Seoul ecosystem
• SNU AI Institute
• KAIST — Beyond AI

The #9 global rank, the #3 Asia rank, and the competitiveness line are Startup Genome’s for GSER 2026. Seoul National University is the source for the AI Institute description. KAIST is the source for the agentic-AI and startup-environment lines. SandBase does not publish a Seoul phone number, a street address, or a city-specific latency SLA on this page.