Ordalin
Submit a tool

Reviewed tool profile

Hyperresearch

Deep-research reports from the web and 90+ data sources, with every citation checked against its source.

Sources checked

Visit website ↗
Pricing
Paid per run, with no subscription: Light costs $9 and Deep $49 per run on the default $0-a-month pay-as-you-go plan. Adding a card makes your first Deep run free (one per new card, once per account); Light runs are charged from the first. Answers MCP calls start at 5¢ with $5 of free credit and no card. Enterprise terms (HIPAA BAA, SSO, private vault) are by conversation. Official source 1 ↗ · Official source 2 ↗
Available as
Web console plus a REST API at api.hyperresearch.ai; reports can be filed into projects and kept in a workspace vault. · Answers MCP server for AI agents and MCP clients, with ask and search tools. Official source 1 ↗ · Official source 2 ↗ Official source 1 ↗
Primary group
Research & Data

Also consider

See all 6 similar tools ↓

What it does

Hyperresearch, from Thalient Labs Inc., is a deep-research service that writes cited reports and checks every citation against its source before delivery. A Deep run plans the question, researches it chapter by chapter across the open web and 90+ data sources such as SEC filings, Census, BLS and PubMed, writes the report in one pass and verifies the cited claims; a Light run takes a shorter path with the same claim check. Answers MCP gives AI agents the same search as short, quoted answers.

Official source 1 ↗ · Official source 2 ↗

Features

  • Citation-checked reports. Every cited figure is read back against the source it cites and corrected when it does not hold; each report ships with its source list and a claims table with the quote behind each claim. Official source 1 ↗ · Official source 2 ↗
  • Deep and Light runs. Light answers one focused question from about 20 to 50 sources in roughly 30 to 40 minutes; Deep investigates a broader question across 150 to 200 sources over 3 to 5 hours. Official source 1 ↗
  • 90+ data sources. Beyond the open web, runs search structured sources such as SEC filings, Census, BLS and PubMed. Official source 1 ↗
  • Source vault. Deep runs save every source to your workspace vault, and later runs search the vault before the web. Official source 1 ↗
  • Answers MCP for agents. Agents get one short answer with the quotes it rests on instead of running their own search loop; calls start at 5¢. Official source 1 ↗
  • Partial reruns. A run can be rerun from a later step at 25%, 50% or 100% of the price, reusing earlier work. Official source 1 ↗

Pricing & free limits

Pay as you go

$0per month

Default plan. Runs are charged as they finish on one monthly invoice, with an optional monthly spend cap. First Deep run free after adding a card. Official source 1 ↗

Light run

$9per run

About 20 to 50 sources, 2,500 to 4,000 words, roughly 30 to 40 minutes. Official source 1 ↗

Deep run

$49per run

About 150 to 200 sources over 3 to 5 hours, with a verification receipt. Official source 1 ↗

Answers MCP

From 5¢per call

ask and search tools for agents, paid from prepaid credit; $5 free credit without a card. Official source 1 ↗

Enterprise

Contact salescontract

HIPAA business associate agreement, private document vault, retention or residency terms, single sign-on or a signed agreement. Official source 1 ↗

Billing details

Runs that fail because of Hyperresearch are refunded automatically, and a run that fails before step 2 completes is not billed. Official source 1 ↗

Reruns are new billed runs priced at a fraction of the run, depending on the step you restart from. Official source 1 ↗

Free access & limits

There is no free plan for reports beyond one free Deep run per new card and account; Light runs are charged from the first. Answers MCP includes $5 of free credit with no card. Official source 1 ↗ · Official source 2 ↗

Best for & limitations

An analyst researching a market or acquisition target gets a sourced report where each number traces to a checked quote. Official source 1 ↗ · Official source 2 ↗

A developer gives an AI agent Answers MCP so it receives short, quoted answers instead of filling its context with search results. Official source 1 ↗

A team running repeated research on one topic builds up a vault that later runs search first. Official source 1 ↗

  • Slow by design. A Light run takes about 30 to 40 minutes and a Deep run 3 to 5 hours; the stated figures are targets, not guarantees. Official source 1 ↗
  • Public pages only. The fetcher stays logged out and does not solve CAPTCHAs; blocked or paywalled sources pause the run until you upload or paste the document yourself. Official source 1 ↗
  • No free tier for reports. Apart from the first Deep run, every report run is paid, and the SDKs are not published to package registries. Official source 1 ↗ · Official source 2 ↗
  • Self-reported benchmarks. The BrowseComp, DeepSearchQA and DRACO comparisons on the homepage are run by Hyperresearch itself. Official source 1 ↗
Hyperresearch product screenshot
Website previewFixed desktop viewport · 1440 × 900

Similar tools

Published tools that share its categories.

DisconfirmAIRead sourced AI news free, then research filings, 13F holdings and standing watches with evidence graded as verified, inferred or doubtful.
Research & DataFree planChecked Oct 15
Visit
XCrawlWeb scraping API that returns LLM-ready Markdown and AI-extracted JSON, with crawl, map and SERP endpoints.
Research & DataFree trialChecked Nov 3
Visit