Skip to main content

FAQ

Frequently asked questions

What ipto.ai is, how retrieval units and pricing work, and how the private-data MCP differs from web search. Still have a question? Get in touch.

The basics

What is ipto.ai? +

ipto.ai is a private-data MCP: agent-first retrieval infrastructure for private, licensed, and proprietary data that agents cannot scrape from the public web. It turns private multimodal data into agent-consumable retrieval units and operates the retrieval, pricing, trust, and audit layer so data owners earn revenue when AI agents use their data.

How is ipto.ai different from web-search MCPs like Exa, Tavily, or Perplexity? +

Web-search MCPs retrieve the public internet. ipto.ai retrieves data behind permission — priced per query, with provenance and audit on every result. It is the private-data counterpart used alongside web search, not a replacement for it. The strongest enterprise agents pair one web-search MCP with ipto.ai for private data.

What is ipto.ai NOT? +

It is not a web-search engine, a scraper, or a vector database. It is the commercial and compliance layer between data owners and the AI agents that consume their data.

How it works

What is a retrieval unit? +

A retrieval unit is a compact, structured data object optimized for agent consumption. It contains facts, entities, dates, obligations, confidence scores, provenance, and pricing metadata — not just raw text. Agents consume units, not blob chunks.

How does provenance work? +

Every ipto.ai result carries provenance: the source document, section, page, a cryptographic hash, and a timestamp. Agents can cite exactly what they used, and data owners can audit every retrieval.

How does pricing work? +

Pricing is usage-based and metered per retrieval, with the per-query price returned inline on every response. No seat licenses. Data owners set their terms and are paid per retrieval; the raw data stays under their control.

How do agents access ipto.ai? +

Through a REST API (api.ipto.ai) and an MCP server, so any MCP-compatible agent — Claude, Cursor, LangGraph, or a custom agent — can register ipto.ai as a tool.

Access, trust & security

How is access controlled? +

Access is enforced at the query layer with scoped API keys and dataset allow-lists — agents see only the datasets they are granted. Every retrieval is metered and logged for a full audit trail.

Is ipto.ai suitable for regulated industries? +

ipto.ai serves financial services, legal, and healthcare use cases where citation-grade accuracy and auditability matter. Access controls, provenance, and audit logs are designed to make proprietary data safe to expose to autonomous systems.

Who is behind ipto.ai? +

ipto.ai is built by IPTO AI Pte. Ltd., incorporated in Singapore and operating globally.

Didn't find your answer?

Reach the ipto.ai team directly — we reply to every legitimate message.