◈ DEVELOPER API

BUILD WITH COUNCIL

Council is an API. Call it from scripts, agents, pipelines, bots — anything that makes decisions.

API QUICK START

curl -X POST https://api.councilzero.com/v1/deliberate \
  -H "Authorization: Bearer $COUNCIL_TOKEN" \
  -d '{
    "question": "Should we approve this PR?",
    "context": "Diff: +142 -38 lines in auth module",
    "mode": "adversarial",
    "max_rounds": 3
  }'

# → { "synthesis": "DECISION: Approve with conditions..." }

INTEGRATION PATTERNS

Real-world systems that call Council when they need a decision.

🤖

OPENCLAW LIFE OS

Your AI assistant consults 101 specialists before acting on complex life decisions. Structured verdicts via Telegram.

OpenClaw → complex decision → Council API → synthesis → Telegram

council_deliberate(question, context, mode="standard")

▸ VIEW DETAILS

🧠

CLAUDE AGENT DELIBERATION

Claude Agent calls Council as a tool when it faces a trade-off. Multi-perspective input for single-agent workflows.

Claude Agent → needs second opinion → council_deliberate() → incorporates synthesis

{"name": "council_deliberate", "input": {"question": "..."}}

▸ VIEW DETAILS

💬

CONVERSATIONAL DECISION BOT

Your Slack/WhatsApp bot escalates complex questions to Council. Simple questions stay fast, complex ones get deliberated.

User message → bot classifies → Council API → synthesis → chat reply

if is_complex(msg): council.deliberate(msg.text)

▸ VIEW DETAILS

💰

TARIFF RENEWAL OPTIMIZER

Before auto-renewing contracts, agents analyze retention risk, margin impact, and competitor pricing.

Cron trigger → customer data → Council API → renew/adjust/flag

council.deliberate(question=f"Renew {customer}?", context=usage_data)

▸ VIEW DETAILS

📝

STUDENT PRACTICE AUDITOR

After grading, Council audits consistency with rubrics and generates per-student feedback reports with action plans.

Grading script → rubric + grades → Council API → consistency report

council.deliberate(question="Consistent?", context=rubric+grades)

▸ VIEW DETAILS

🔀

PR REVIEW PIPELINE

GitHub Actions triggers a Council deliberation on every PR. Agents review architecture, security, and test coverage.

PR opened → CI extracts diff → Council API (adversarial) → PR comment

council.deliberate(question="Review PR", mode="adversarial")

▸ VIEW DETAILS

🎯

LEAD QUALIFICATION

CRM webhook fires on new leads. Council evaluates ICP fit and recommends the right plan tier.

New lead → CRM webhook → Council API → hot/warm/cold + plan

council.deliberate(question=f"Qualify {lead}", context=icp)

▸ VIEW DETAILS

📊

DATA PIPELINE QA

After ETL completes, agents deliberate on data quality, anomalies, and drift. Approve or halt downstream.

ETL done → quality metrics → Council API → approve/alert/block

council.deliberate(question="Data OK?", context=metrics)

▸ VIEW DETAILS

🚨

INCIDENT TRIAGE

Critical alerts get deliberated before paging humans. Agents assess real severity, blast radius, and root cause.

Alert → PagerDuty webhook → Council API → triage report

council.deliberate(question=alert.summary, mode="standard")

▸ VIEW DETAILS

🛡

CONTENT MODERATION

Flagged content gets adversarial deliberation: context, intent, precedent. More nuanced than rule engines.

Content flagged → Council API (adversarial) → verdict + confidence

council.deliberate(question="Harmful?", mode="adversarial")

▸ VIEW DETAILS

MCP & AGENT TOOLS

COMING SOON

Council as a first-class tool in your AI agent's toolkit

Any AI agent that supports MCP can call Council for multi-perspective deliberation. Your agent gets a second opinion — from 101 specialized agents — before making critical decisions.

mcp_config.json

{
  "mcpServers": {
    "council": {
      "command": "council-mcp",
      "args": ["--ecosystem", "my-council"],
      "env": { "COUNCIL_TOKEN": "your-token" }
    }
  }
}

START BUILDING

API access from Pro ($5/mo)

▸ JOIN THE WAITLIST ▸ VIEW USE CASES

See Use Cases →

Start free →

◎Partners ◉Showcase →Waitlist