CivicGraph
Agent API
Structured intelligence on Australian government spending, procurement, political donations, and community organisations. Built for AI agents that need real data, not hallucinated answers.
Quick Start
POST /api/agent
Content-Type: application/json
Authorization: Bearer cg_live_...
{
"action": "search",
"query": "Commonwealth Bank"
}GET /api/agent
API Keys
Try It Now
Available Actions
Entity Search
"action": "search"Find entities by name or ABN. Returns power scores, cross-system presence, and dollar flows.
{
"action": "search",
"query": "Commonwealth Bank"
}Array of entities with gs_id, canonical_name, entity_type, power_score, system_count, total_dollar_flow
Entity Profile
"action": "entity"Full entity intelligence — power score, 7-system presence, board members, procurement history, political donations, justice funding.
{
"action": "entity",
"abn": "48123123124"
}Entity profile + board member list with roles and appointment dates
Power Index
"action": "power_index"Top entities ranked by cross-system power score. Spans procurement, justice, donations, charity, foundation, evidence, and tax systems.
{
"action": "power_index",
"limit": 10,
"min_systems": 3
}Ranked entities with power_score, system_count, and dollar breakdowns per system
Funding Deserts
"action": "funding_deserts"Most underserved local government areas in Australia. Scored by disadvantage (SEIFA), remoteness, and funding shortfall.
{
"action": "funding_deserts",
"state": "NT",
"limit": 10
}LGAs ranked by desert_score with IRSD decile, entity counts, and dollar flows
Revolving Door
"action": "revolving_door"Entities with multiple influence vectors — lobbying, political donations, and government contracts. Cross-referenced automatically.
{
"action": "revolving_door",
"limit": 10
}Entities with revolving_door_score, influence vectors, donation and contract totals
Natural Language Query
"action": "ask"Ask any question in plain English. CivicGraph generates SQL, executes it across all datasets, and returns structured results with an AI explanation.
{
"action": "ask",
"query": "How much does QLD spend on youth justice?"
}Query results as structured data + AI explanation + generated SQL
Agent Use Cases
Procurement Agent
Evaluate Australian suppliers before bid decisions. Check cross-system presence, contract history, political donation patterns, and compliance signals.
Policy Research Agent
Analyse funding flows, identify service gaps, compare state spending on youth justice, disability, and Indigenous services.
Due Diligence Agent
Entity profile, board interlock detection, revolving door analysis. Automated background checks on government contractors.
Compliance Agent
Monitor entity changes, new contracts, board appointments. Detect conflicts of interest and emerging interlock patterns.
Data Sources
Pricing
- 20 requests/minute
- All 6 actions
- IP-based rate limiting
- Community support
- 60 requests/minute
- Usage dashboard + analytics
- NL→SQL queries
- Up to 5 API keys
- Unlimited requests
- Dedicated support
- Custom endpoints
- White-label + SSO
Works With
claude mcp add civicgraphPOST /api/agentCivicGraphTool()curl, fetch, axiosBuilt For Real Work
“We mapped $74B in federal procurement flows to identify conflict-of-interest patterns that would take a human analyst months.”
“Cross-referencing political donations with government contracts across 560K entities — automatically. No hallucinations, just data.”
“Identified 1,582 funding deserts — local government areas where disadvantage is high but funding is almost zero.”
“1,155 evidence-based interventions from ALMA, linked to delivery organisations. Know what works and who delivers it.”
Why CivicGraph for Agents
General-purpose LLMs hallucinate government data. They don't know who holds what contracts, which entities donate to which parties, or where funding deserts exist. CivicGraph does.
Every response comes from structured, cross-referenced datasets — not generated text. 560K entities linked by ABN across 11 data sources. Every number is auditable.
An agent that buys a CivicGraph query for $0.01 gets a verified answer. An agent that tries to answer from training data gets a plausible-sounding wrong answer. In procurement and compliance, the difference matters.