AiMatch

Hyperresearch

Hyperresearch produces research reports with sources, supporting quotes, and citation checks for researchers and teams. Its vault, MCP server, and REST API let users and agents reuse evidence across questions and workflows.

Visit Hyperresearch Compare alternatives Own this tool? Claim it
Hyperresearch screenshot

Who it's for

  • Research and strategy teams that need reports with source lists, claim quotes, and citation-check receipts.
  • Developers building agent workflows who want MCP access to a research vault or Answers MCP responses grounded in quotes.
  • Teams reviewing documents before publication that need citations, quoted passages, and retractions checked against sources.

Not the right fit if…

  • Teams that need a synchronous REST research response: the page says runs are asynchronous and reports are available after a webhook signals completion.
  • People who need a free first Deep run without providing payment details: the page says an account and card are required for that offer.

How it fits your workflow

Submit a research question and choose a Light or Deep run. Hyperresearch searches the vault first, then gathers web pages and structured data, investigates the evidence, writes the report, and checks cited claims against their sources.

Review the report alongside its source list and claim-and-quote table; Deep reports also include a verification receipt. Sources and interim work are saved in the vault for later searches.

For agent workflows, authorize a client such as Claude, Cursor, or ChatGPT to use the remote MCP server. For pipelines, start a run through the REST API, then retrieve the report after the completion webhook.

Pricing

Free: The page offers $5 in free Answers MCP credit without a card. It also says the first Deep run is free after creating an account and adding a card.

Light

$9 per run

  • · About 30–40 minutes
  • · Answer-first report of about 2,500–4,000 words, sized to the question
  • · Cited claims checked against their sources before delivery

Deep

$49 per run

  • · About 3–5 hours
  • · Research across about 150–200 sources
  • · Verification receipt with citation-check findings

Answers MCP

Asks from $0.05; searches from $0.05

  • · $5 free credit, no card needed
  • · Short answers checked against word-for-word quotes or a plain “not found”

Prices checked on Nov 2, 2026 from the vendor's site. They can change; confirm before you buy.

Key features

Evidence-backed research reports
Light and Deep runs research a question and produce reports with a source list and a table of claims and supporting quotes. Deep runs use a multi-stage process that includes research chapters, evidence mapping, report planning, and claim checking.
Citation verification receipts
The verification process checks cited claims against their sources, records per-citation verdicts, checks quote integrity, and flags retracted papers. Deep reports include a verification receipt.
Searchable research vault
Sources, interim notes, claims, and other run artifacts are kept in a workspace. Users can search by words or meaning, and subsequent runs search the vault before the web.
Answers MCP
Agents can ask a question and receive a short answer checked against word-for-word quotes, or a plain “not found” response. The search tool can also return ranked results from the open web and structured sources.
Structured source access
Research can draw on more than 40 data APIs, including Census, BLS, BEA, SEC EDGAR, OpenAlex, and PubMed, as well as public web pages and records.
Document citation checking
Users can submit a document and its sources for citation verdicts, quote checks, retraction checks, and checks for sources that are effectively the same story.

Works with

  • Claude, Cursor, and ChatGPT
  • Custom agents
  • REST API and webhooks
  • Structured data APIs and sources
  • Parallel, Exa, and Valyu

Limitations to know

  • API research runs are asynchronous; the page says a run returns an ID first and the completed report is retrieved after a webhook indicates it is ready.
  • The first Deep run requires an account and a card, even though the run itself is offered free.
  • The stated Light and Deep run durations are about 30–40 minutes and 3–5 hours, respectively.

Alternatives to consider

  • Ressearch AI

    Choose it if your scientific research workflow needs literature search combined with Python/R analysis, visualization, and publication-ready exports.

  • Tavily (gpt-researcher)

    Choose it if your primary need is a production API for agents to retrieve, extract, and crawl current web information rather than run full evidence-checked reports.

  • WebSearchAPI.ai

    Choose it if you need a web search and retrieval API that returns results, structured data, or grounded responses from multiple search engines and public communities.

Compare Hyperresearch side by side

Getting started

Create an account. The first Deep run is offered free after adding a card; MCP authorization takes place through a browser screen where users select a workspace and permissions.

  1. Create an account on Hyperresearch.
  2. Choose a Light or Deep run and submit a research question; add a card if using the first-free-Deep-run offer.
  3. Review the report, source list, and claim-and-quote table; check the verification receipt for a Deep run.
  4. Optionally authorize an MCP client or use the REST API for agent and pipeline workflows.

Frequently asked questions

What do I receive from a research run?

The page says a run returns a report, source list, and a table of claims with their supporting quotes. A Deep report also includes a verification receipt.

What is the difference between Light and Deep?

Light is described as a roughly 30–40-minute run producing an answer-first report of about 2,500–4,000 words. Deep takes about 3–5 hours, uses about 150–200 sources, and includes a verification receipt.

Can I connect my own AI agent?

Yes. The page describes a remote MCP server that can be authorized from clients such as Claude, Cursor, and ChatGPT, as well as from an agent you wrote.

Does Hyperresearch provide an API?

Yes. The page describes a REST API with scoped keys for pipelines. Runs are asynchronous: the API returns a run ID, and a webhook signals when the report is ready.

Can Hyperresearch check a document I already have?

Yes. The page describes a verification API that checks citations against sources, quote integrity, retracted papers, and sources that may be the same story.