Fodda API & MCP — Expert Intelligence for AI Systems
Connect your AI systems, agent pipelines, and web applications to 300+ curated knowledge graphs, 100+ live institutional feeds, Human Practitioner Agents, and factual claim verification via REST API, Model Context Protocol (MCP), and Agent-to-Agent (A2A) machine protocols.
Core API Surface & Machine Contract
This reference documents Fodda's Core API Families. For the exhaustive, machine-readable specification of all 150+ platform operations, parameter constraints, and schemas, inspect the Fodda OpenAPI 3.1 Specification.
Connection Methods
- REST API:
https://api.fodda.ai/v1— Direct deterministic HTTP integration. - MCP Server (HTTP):
https://mcp.fodda.ai/mcp— Streamable HTTP for Claude, Cursor, and IDEs. - Tokenized MCP:
https://mcp.fodda.ai/c/<token>— Streamable transport with pre-authenticated connection token. - ChatGPT MCP:
https://mcp.fodda.ai/chatgpt— OAuth 2.1 authentication for custom ChatGPT Connectors. - Copilot Studio MCP:
https://mcp.fodda.ai/copilot— Curated 17-tool MCP profile for Microsoft Copilot Studio. - A2A & Machine Settlement: Autonomous delegation via
/.well-known/agent-card.jsonwith HTTP 402 MPP challenge.
Authentication
- Account API Key:
X-API-Key: sk_live_...orAuthorization: Bearer sk_live_... - Autonomous MPP / SPT: Unauthenticated requests return HTTP 402 with price challenge; settle via
Authorization: Bearer spt_...(Direct REST only). - Enterprise OIDC:
Authorization: Bearer <OIDC_JWT> - Private Connection Token:
https://mcp.fodda.ai/c/<token>for header-incapable clients (treat the complete URL as a private secret).
Quickstart: Search a Knowledge Graph
Example request querying the Retail knowledge graph for AI shopping agent trends:
cURL
curl -X POST "https://api.fodda.ai/v1/graphs/retail/search" \
-H "Authorization: Bearer $FODDA_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"query": "How are AI shopping agents changing retail checkout?",
"limit": 5
}'
Python
import os, requests
res = requests.post(
"https://api.fodda.ai/v1/graphs/retail/search",
headers={"Authorization": f"Bearer {os.environ['FODDA_TOKEN']}", "Content-Type": "application/json"},
json={"query": "How are AI shopping agents changing retail checkout?", "limit": 5},
timeout=30
)
res.raise_for_status()
print(res.json())
JavaScript
const res = await fetch("https://api.fodda.ai/v1/graphs/retail/search", {
method: "POST",
headers: {
"Authorization": "Bearer " + process.env.FODDA_TOKEN,
"Content-Type": "application/json"
},
body: JSON.stringify({ query: "How are AI shopping agents changing retail checkout?", limit: 5 })
});
const data = await res.json();
console.log(data);
Representative Response JSON
{
"ok": true,
"dataStatus": "TREND_MATCH",
"meta": {
"query": "How are AI shopping agents changing retail checkout?",
"graph_id": "retail",
"decision": "ANSWER",
"execution_time_ms": 142
},
"rows": [
{
"trend_id": "trend-agentic-checkout-602",
"name": "Autonomous Agentic Commerce",
"summary": "AI agents increasingly execute end-to-end shopping transactions, bypassing traditional web front-ends.",
"evidence": [
{
"article_id": "art-9481",
"title": "Retail AI Protocol Benchmarks",
"source_url": "https://psfk.com/2026/08/retail-ai-benchmarks.html",
"published_at": "2026-08-14"
}
]
}
]
}
Core Endpoints Summary
| Method & Path | Description | Access |
|---|---|---|
GET /v1/graphs/catalog | Live catalog of 300+ domain, report, and expert graphs | Public |
GET /v1/catalog.okf | Open Knowledge Format bundle for AI agents | Public |
GET /v1/graph-slice?graph_id=retail | Sample JSON-LD data slice | Public |
GET /v1/experts/search | Search 4,150+ specialist roster | Public |
POST /v1/graphs/{graph_id}/search | Hybrid vector and keyword search within a specific graph | Authenticated |
POST /v1/human-agents/consult | Consult living industry figures' digital twins | Authenticated |
POST /v1/analysts/consult | Consult synthetic domain analysts | Authenticated |
POST /v1/brand-intelligence/{brand} | Cross-graph brand footprint, signals, and trends | Authenticated |
POST /v1/research/deep-dive | Autonomous multi-pass deep research synthesis | Authenticated |
POST /v1/verify/claim | 5-gate anti-hallucination factual claim verification | Authenticated |
POST /v1/supplemental/context | Real-time macroeconomic feeds across 100+ sources | Authenticated |
Operational Error Contracts
All API errors return a standard JSON envelope with a machine-readable code, clear message, and correlation request ID:
{
"error": {
"code": "invalid_parameter",
"message": "The specified graph_id 'unknown_graph' was not found in the active catalog.",
"request_id": "req_8f1a92bc4e02",
"retryable": false
}
}
- 400 Bad Request: Invalid parameter, malformed JSON, or missing required field (e.g.
INVALID_ARGUMENT). - 401 Unauthorized: Missing or invalid API key or connection token.
- 402 Payment Required: MPP Challenge for unauthenticated autonomous requests; returns price challenge and settlement address.
- 404 Not Found: Graph ID, node ID, or analyst name not found in the active catalog.
- 429 Rate Limit Exceeded: Request ceiling reached. Emits
Retry-Afterheader with backoff seconds. - 500 Server Error: Transient upstream or database error. Client should retry with exponential jitter.
Data Handling, Privacy & Retention
In accordance with Fodda's Compliance Dossier: query text is retained for 12 months solely for billing reconciliation and telemetry, after which it is purged. Result text is never retained. Client prompts and responses are never used for model training. Contractual zero-retention query logging is available for enterprise tiers.
Pricing & Quotas
Published USD rate: 50¢ per API call. Fodda Base plan includes 100 free calls every month. Daily burst ceiling without card on file: 50 calls/day. Commercial deliverables (Brand Intelligence $30.00, Deep Research $100.00) are billed at flat published rates.