Threads • Attachments • Decisions • Follow-ups

Most “work context” lives in email - long threads, scattered approvals, buried attachments.

iGPT is an API that turns that email context into reliable, automation-ready output:

• Summaries you can trust
• Action items you can run
• Decisions you can log
• Follow-ups you can trigger

• ✅ Action items with owner + due_date (schema-enforced)
• 🧾 Decision logs + approval timelines across threads
• 🚨 Risk & blocker detection (what’s stuck, who’s waiting)
• 📎 Attachment-aware answers (PDFs, docs, spreadsheets referenced in email)
• 🔁 Follow-up automation (missing replies, stale threads, next steps)

npm install igptai

import IGPT from "igptai";

const igpt = new IGPT({ apiKey: process.env.IGPT_API_KEY });

// Connect a datasource (per user)
const auth = await igpt.connectors.authorize({
  service: "spike",
  scope: "messages",
  user: "user_123"
});

console.log("Authorize:", auth.url);

// Recall: ask with email context
const res = await igpt.recall.ask({
  user: "user_123",
  input: "List open action items from this week's email threads. Include owner + due date."
});

console.log(res);

• Node.js SDK: npm install igptai
• Python SDK: pip install igptai

• Schemas, not prompt parsing → get strict JSON output you can execute
• Cited context → audit and verify answers back to source messages/files
• Streaming support → build live UIs and agent pipelines
• No-throw SDK design → errors returned as { error: string }

• Website: https://www.igpt.ai/
• Docs: https://docs.igpt.ai/
• Playground: https://igpt.ai/hub/playground/
• Pricing: https://www.igpt.ai/pricing/
• Book a demo: https://www.igpt.ai/contact-sales/
• Contact: hello@igpt.ai