Main Site Book a Demo
DOCUMENTATION

INGOUDE Platform

The complete reference for deploying, managing, and scaling enterprise autonomous AI operations — from first agent to production fleet.

ℹ️
Platform v2.4 is now live

Multi-agent orchestration, streaming responses, and improved latency. See Changelog.


Quickstart Guide

Get from zero to a running AI agent in five steps. Prerequisites: Node.js 18+ or Python 3.10+.

  1. 1

    Install the INGOUDE CLI

    Available via npm and pip.

bash
npm install -g @ingoude/cli
# or: pip install ingoude-cli
  1. 2

    Authenticate

    Generate your key from the Dashboard.

bash
ingoude login --key ing_live_xxxxxxxxxxxxxxxxxxxx
  1. 3

    Create an Agent

    Choose a template matching your use case.

bash
ingoude agents create --name "my-agent" --template revenue-intel --region us-east-1
  1. 4

    Link a Data Source

    Connect Salesforce, Stripe, or any CRM/ERP.

bash
ingoude sources connect salesforce --agent my-agent --auth-token $SF_TOKEN
  1. 5

    Deploy & Monitor

    Start and open the Operations Dashboard.

bash
ingoude agents start my-agent
# ✓ Agent spawned · dashboard: https://app.ingoude.ai/agents/my-agent

Security & Compliance

INGOUDE is SOC 2 Type II certified with continuous compliance monitoring.

  • mTLS Authentication — Every node verifies identity via mutual TLS certificates.
  • Data Locality Controls — Restricted agents that never send data outside your VPC region.
  • PII Auto-Scrubbing — Automatic redaction before context reaches the reasoning engine.
  • RBAC & Audit Logs — Granular role-based access with immutable audit trails.
  • Zero-Knowledge Architecture — INGOUDE never stores your proprietary data.

CORE PLATFORM
🤖
PRODUCT

AI Agent Hub

The centralized control plane for your entire autonomous AI workforce.

🔁

Reasoning Loop Engine

Think → Plan → Act cycles with full auditability of every reasoning step.

📊

Capacity Autoscaling

Spawn parallel instances during event spikes. Scale from 1 to 10,000 agents in seconds.

🧠

Memory Persistence

Agents retain context via vector stores — episodic, semantic, and procedural memory layers.

🔗

Multi-Agent Orchestration

Define agent graphs where specialized agents collaborate and hand off tasks.

Agent Configuration Schema

JSON · agent-config.json
{ "agent": { "name": "revenue-intel-primary", "template": "revenue-intel",
    "replicas": 3, "reasoning": { "model": "ingoude-reasoning-v2", "max_iterations": 25 },
    "tools": ["salesforce", "stripe", "slack-notify"] } }
🏢
PRODUCT

Enterprise AI Solutions

The complete enterprise AI transformation suite — strategy, implementation, and managed operations.

Custom Model Fine-Tuning

Python
import ingoude
client = ingoude.EnterpriseClient(api_key="ing_live_...")
job = client.fine_tuning.create(base_model="ingoude-reasoning-v2",
    training_file="s3://my-bucket/training-data.jsonl", epochs=3)
⚙️
PRODUCT

Autonomous Operations

End-to-end workflows that run continuously with configurable approval gates.

YAML · workflow.yml
name: invoice-reconciliation
trigger: { type: schedule, cron: "0 9 * * 1-5" }
steps:
  - { name: fetch-invoices, tool: quickbooks.list_invoices, params: { status: overdue } }
  - { name: analyze-risk, agent: risk-scoring-agent }
  - { name: escalate, tool: slack.send_message, condition: "{{ score }} > 0.8" }
📺
PRODUCT

AI Operations Dashboard

Real-time observability — performance metrics, reasoning traces, cost analytics.

TypeScript
const metrics = await client.analytics.getFleetMetrics({
  window: '24h', granularity: 'hourly',
  include: ['token_usage', 'decision_count', 'cost_usd'],
}); // { total_decisions: 14823, avg_latency_ms: 342, cost_usd: 18.40 }

INTELLIGENCE
💹
PRODUCT

Revenue Intel

AI agents trained on CRM, ERP, and payment data to identify churn signals with 98% accuracy.

JSON · Signal Config
{ "data_sources": ["salesforce", "stripe", "hubspot"],
  "signals": { "churn_risk": { "threshold": 0.72, "action": "alert-csm" },
               "expansion":  { "threshold": 0.85, "action": "create-opportunity" } } }
🧩
PRODUCT

Enterprise Decision Intelligence

Transform complex decisions into structured, auditable AI reasoning processes.

Python
response = client.decisions.evaluate(context="Applicant: Acme Corp, ARR $4.2M...",
    policy="enterprise-credit-policy-v3")
print(response.recommendation) # "APPROVE" | "REVIEW" | "DECLINE"
💳
PRODUCT

AI Credit Underwriting

Process hundreds of data points in seconds — bureau data, bank data, tax docs, business financials.

🏦

Bureau Data

Equifax, Experian, TransUnion with automated tri-merge logic.

🔗

Bank Connectivity

Plaid, Finicity, MX for 90-day cash flow analysis.

📋

Tax Documents

IRS 4506-C. Supports 1040, W-2, 1099 forms.

🏢

Business Financials

P&L, balance sheet, A/R aging via QuickBooks or PDFs.

⚠️
PRODUCT

Risk Scoring Agents

Continuous risk assessment — counterparty, operational, market, and regulatory risk.

cURL
curl -X POST https://api.ingoude.ai/v1/risk/score \
  -H "Authorization: Bearer $INGOUDE_API_KEY" \
  -d '{"entity_id":"acct_9xZpQ2","risk_dimensions":["counterparty","market"]}'
💰
PRODUCT

Pricing Optimization Agents

Dynamic pricing that continuously analyzes competitive data and demand signals.

JSON · Optimization Config
{ "strategy": "revenue_maximization",
  "constraints": { "min_margin_pct": 22, "competitor_index": "within_5_pct" },
  "update_frequency": "6h" }

INDUSTRY SOLUTIONS
🛡
INDUSTRY

Compliance & Governance

Continuous monitoring across SOX, GDPR, HIPAA, Basel III, EU AI Act, and 18+ frameworks.

🏦

Financial

SOX, Basel III, MiFID II, Dodd-Frank, AML/KYC, PCI DSS.

🔒

Data Privacy

GDPR, CCPA, PIPEDA, LGPD, PDPA — 18 jurisdictions.

🏥

Healthcare

HIPAA, HITECH, FDA 21 CFR Part 11, ISO 27001.

🌍

AI Governance

EU AI Act, NIST AI RMF, ISO 42001 readiness.

🏥
INDUSTRY

Healthcare Compliance

HIPAA monitoring, PHI detection, and clinical documentation automation.

cURL · PHI Detection
curl -X POST https://api.ingoude.ai/v1/hipaa/detect-phi \
  -H "Authorization: Bearer $INGOUDE_API_KEY" \
  -d '{"text":"Patient John D., DOB 1980-04-12, MRN 48291..."}'
🏨
INDUSTRY

Hospitality Intelligence

Revenue management, guest personalization, and demand forecasting for hotels and travel.

  • Dynamic rate recommendations updated every 15 minutes based on demand signals
  • Competitor rate parity monitoring across Booking.com, Expedia, direct channels
  • Guest lifetime value scoring for loyalty and upsell prioritization

API REFERENCE

Authentication

All API requests require a Bearer token in the Authorization header.

HTTP
GET /v1/agents HTTP/1.1
Host:          api.ingoude.ai
Authorization: Bearer ing_live_xxxxxxxxxxxxxxxxxxxx
Content-Type:  application/json

API Key Scopes

Scope Type Description
agents:read Read List and retrieve agent status
agents:write R/W Create, update, delete agents
analytics:read Read Access fleet metrics and cost data
admin Full Manage users, billing, org settings

Agent Endpoints

Method Endpoint Description
GET /v1/agents List all agents
POST /v1/agents Create a new agent
GET /v1/agents/{id} Retrieve agent config and status
POST /v1/agents/{id}/spawn Initialize a reasoning loop
GET /v1/agents/{id}/traces Retrieve reasoning trace history
DELETE /v1/agents/{id} Delete an agent

Webhooks

Subscribe to agent events. INGOUDE sends signed HTTP POST requests to your endpoint when events occur.

JSON · Webhook Payload
{ "type": "agent.decision_ready", "data": {
    "agent_id": "agt_revenue-intel", "recommendation": "EXPAND", "confidence": 0.91 } }

Rate Limits

Starter
500
req / min
Professional
2,000
req / min
Enterprise
Custom
negotiated

Error Codes

Status Code Meaning
400 invalid_request Malformed body or missing required field
401 unauthorized Invalid or expired API key
403 forbidden Key lacks required scope
404 not_found Resource does not exist
429 rate_limit_exceeded Too many requests — check Retry-After

TECHNICAL

SDK Reference

🐍

Python SDK

Full async support. Python 3.10+. pip install ingoude

📦

TypeScript SDK

ESM + CJS. Zod-validated responses. npm i @ingoude/sdk

Java SDK

Spring Boot autoconfiguration. Java 17+.

🐹

Go SDK

Idiomatic Go with context propagation.

Python · Streaming
async with client.agents.stream("agt_revenue-intel") as stream:
    async for chunk in stream:
        print(chunk.thought) # Real-time reasoning steps

Architecture Overview

System Diagram
  ┌──────────────────────────────────────────────────┐
  │              CLIENT LAYER                        │
  │  Web App · REST API · SDKs · CLI · Webhooks      │
  └──────────────────┬───────────────────────────────┘
                     │
  ┌──────────────────▼───────────────────────────────┐
  │              API GATEWAY                          │
  │  Auth / mTLS · Rate Limiting · Request Routing   │
  └────────┬─────────────────┬────────────────────────┘
           │                 │
  ┌────────▼───────┐  ┌──────▼──────────────────────┐
  │  AGENT MESH     │  │    REASONING ENGINE           │
  │  Orchestration │◄►│  Think → Plan → Act          │
  └────────┬───────┘  └──────────────────────────────┘
           │
  ┌────────▼───────────────────────────────────────────┐
  │         DATA & INTEGRATION LAYER                    │
  │  Salesforce · Stripe · SAP · Vector Store          │
  └────────────────────────────────────────────────────┘

Deployment Guide

VPC Deployment via Helm

bash · Helm
helm repo add ingoude https://charts.ingoude.ai
helm install ingoude-runtime ingoude/agent-runtime \
  --set apiKey=$INGOUDE_API_KEY --namespace ingoude-system --create-namespace

Integrations

  • CRM: Salesforce, HubSpot, Microsoft Dynamics 365, Pipedrive
  • Finance: Stripe, QuickBooks, NetSuite, SAP S/4HANA
  • Comms: Slack, Teams, Jira, Linear, Google Workspace

Changelog

v2.4 — March 2026

  • Multi-agent orchestration — Agent graphs with typed handoffs and shared memory
  • Streaming responses — Real-time reasoning token streams via SSE
  • EU AI Act compliance mode — Auto documentation for high-risk AI deployments

v2.3 — January 2026

  • Healthcare Compliance (HIPAA PHI detection) — GA
  • Pricing Optimization Agents: 15-minute update cycle
  • Python SDK v2 with full async support