Eureka Simulations Agentic API
Hosted MCP · Streamable HTTP · Bearer eak_

Point an agent at Eureka & let it run a live simulation.

Author a business simulation from a spec, deploy it to a cohort, hand every team its one-click login, and drive the rounds — from Claude Code, Cursor or your own agent. The same key also manages the cohort and reads the rubric scores.

claude / cursor — mcp config
// self-discovers tools filtered to your key's scopes
{
  "mcpServers": {
    "eureka": {
      "command": "npx",
      "args": ["-y", "mcp-remote",
        "https://eurekasimulations.com/_api/mcp/server.php",
        "--header", "Authorization: Bearer ${EUREKA_API_KEY}"],
      "env": { "EUREKA_API_KEY": "eak_…" }
    }
  }
}
16
tools across 4 categories
Free
reads, session state & team links
Live
deploy, autologin, advance rounds
0→4
usage tiers, auto-promoted on spend
Get started

Author, then run

All tools — authoring, the live-session family (deploy, team links, round control), programs and reads — connect over the hosted MCP server; tools/list returns exactly what your key's scopes allow. Every tool is also callable as raw bearer REST.

terminal
# add the connector to Claude Code (author + cohort + read tools)
claude mcp add --transport http eureka \
  https://eurekasimulations.com/_api/mcp/server.php \
  --header "Authorization: Bearer $EUREKA_API_KEY"
# deploy a sim to a cohort, then advance the round for every team
curl -s https://eurekasimulations.com/_api/mcp_express_session.php \
  -H "Authorization: Bearer eak_…" \
  -d '{"externalid":"42","num_teams":6,"name":"MBA cohort A"}'

curl -s https://eurekasimulations.com/_api/mcp_express_session_state.php \
  -H "Authorization: Bearer eak_…" \
  -d '{"sessionid":5012,"action":"next_round"}'
# who am I? (free) — every tool is also raw REST
curl -s https://eurekasimulations.com/_api/mcp_whoami.php \
  -H "Authorization: Bearer eak_…"

# hand each team its one-click autologin link
curl -s https://eurekasimulations.com/_api/mcp_express_session_links.php \
  -H "Authorization: Bearer eak_…" -d '{"sessionid":5012}'
Trust model

Scoped, metered, and yours to steer

A key carries only the scopes you grant, spends transparently in credits, and is audited on every call. Facilitation stays under human control — the round only advances when someone (or their agent, on their behalf) says so.

Credits & tiers

Reads, session state and team-link fetches are free. Authoring a sim or opening a program costs a few credits. Accounts start on Free and auto-promote (Tier 1–4) as cumulative spend grows, lifting rate limits and monthly caps.

  • 0Reads, session state, team autologin links
  • 2–5Author / update a simulation, open a program
  • 10Delete a program (destructive)

What a key can't do

Guardrails are enforced server-side, not by the agent's good behavior:

  • scopeOnly the tools you granted are callable
  • ownerSessions & programs are bound to the key owner
  • HITLDestructive actions (delete, publish) require an approval token
  • auditEvery call logged with scope, cost, and outcome
1

Deploy the session

express_session_create spins up a live game for N teams and returns each team's autologin link to hand out.

2

Facilitate the rounds

express_session_state with play / pause / next_round drives the whole cohort; poll state to watch teams decide.

3

Read the scores

read_frameworks returns the rubric plus each team's AI & faculty evaluation for a round — the input to a debrief.

Reference

The tool catalog

All 16 tools for authoring and running Eureka Express simulations. Programs and reads are shared with Pro Evaluation System — the same key spans both.

Every tool is discoverable via tools/list and callable as raw bearer REST.
Filter