iGPT

iGPT is an enterprise-grade email intelligence API that transforms the unstructured data within email communications and attachments into structured, actionable intelligence for AI agents and automated workflows. It solves a critical gap where most AI systems fail: reliably accessing and understanding the complex context of email threads, documents, and historical conversations. Designed for developers and enterprises, iGPT provides a secure, unified gateway to deep email context through a single API call. This eliminates the immense engineering burden of building and maintaining custom pipelines for parsing, chunking, indexing, and prompt tuning. By delivering trusted, context-aware answers with full citations, iGPT enables businesses to confidently deploy intelligent automation for sales, customer support, operations, and compliance, turning the chaos of email into a powerful, structured data source.

Unified Contextual Endpoint

Replace complex multi-step retrieval systems with a single natural language API request. iGPT's unified endpoint seamlessly handles retrieval, context shaping, and reasoning in one integrated pipeline, delivering answers ready for automation without any prompt engineering or manual pipeline management required.

Deep Email & Attachment Processing

iGPT automatically and deeply indexes all email data, including complex threads with multiple participants. It also processes attachments, extracting text, data, and structure from documents, PDFs, and spreadsheets to provide a complete view of the conversation's context and content.

Hybrid Retrieval with Citations

The system employs a sophisticated hybrid retrieval method combining semantic, keyword, and filtered search, which is then scored and re-ranked in one call. Every answer is returned with direct citations back to the source emails and attachments, ensuring verifiability and trust.

Enterprise-Grade Security & Auditability

Built with strict security principles, iGPT offers zero data training, OAuth-only processing with Role-Based Access Control (RBAC), and zero data retention. Each inference is processed in memory, and a full audit trail ensures all actions are traceable, keeping data under your company's control.

Intelligent Email Assistants

Build AI agents that can draft, prioritize, summarize, and act on emails with a complete understanding of the entire thread history, participant roles, and attached documents, dramatically improving productivity and response quality.

Customer Support Copilots

Empower support teams with agents that instantly reconstruct the full customer story across long, messy email chains. These copilots can understand tone shifts, track issues, and reference attachments to provide accurate, context-rich support.

Workflow & CRM Automation

Automatically turn email conversations into structured data for other systems. Extract action items, deadlines, approvals, and deal momentum to create tasks in project tools or update records in CRM platforms like Salesforce, directly from email threads.

Compliance & Audit Trail Generation

Automate the tracking of critical communications for legal and regulatory compliance. iGPT can trace feedback, decisions, approvals, and rationale back to the original email and attachment sources, creating a reliable and searchable audit trail.

How does iGPT handle data privacy and security?

iGPT is built on a foundation of enterprise-grade security. It practices zero data training, meaning your data is never used to train models. It uses OAuth-only processing with RBAC, enforces zero data retention (processing happens in memory), and provides a full audit trail, ensuring data never leaves your control.

What makes iGPT different from using a generic LLM on email text?

Generic LLMs lack the specialized understanding of email structure, threading, and attachment context. iGPT provides a complete pipeline that deeply indexes, reconstructs threads, processes attachments, and performs optimized retrieval to deliver accurate, cited answers from messy email data through one simple call, eliminating months of custom engineering.

Can iGPT process data from multiple email sources?

Yes, iGPT is designed to connect to and unify data from various enterprise email sources. It deeply indexes emails and attachments from these connected sources, allowing a single API query to retrieve context from across all of them, creating a unified intelligence layer.

What is the Context Engineering Framework (CEF)?

The Context Engineering Framework (CEF) is iGPT's performance and quality tier system. It allows developers to choose the optimal balance of speed, cost, and reasoning depth for their specific use case by selecting different CEF tiers (e.g., cef-1-normal) when making an API request.