Agentic deep research platformLive

Clarixity

Autonomous research agents. Every claim cited and verified.

Clarixity deploys a team of AI agents that plan the investigation, search the open web and your own documents, read a hundred sources in minutes and deliver analyst grade reports where every claim is cited and checked. Run a question once, or put it on a schedule and track how the answer moves.

Open ClarixityAll products
Research depths
5Research depthsFrom a one credit answer to an exhaustive report.
Built in briefs
7Built in briefsCompetitive analysis, literature reviews, due diligence and more.
Every citation
CheckedEvery citationEach cited sentence is checked against its source after writing.
The Clarixity landing page in signal blue on black, with the headline Autonomous research agents, every claim cited and verified, and a live answer beside it.

Capabilities

What Clarixity does

  • 01

    Citations you can check

    Every claim carries a small superscript number. Hovering one shows the sentence from the source that supports it, and wide screens keep a source rail and a contents list beside the answer.

  • 02

    Claims checked after writing

    Each cited sentence is checked against its sources. A sentence the sources do not support gets an unverified mark, and facts the sources contest get a Sources disagree panel under the lead.

  • 03

    Source quality badges

    Sources are marked as peer reviewed, government, primary, news or opinion, and an old one is flagged as possibly outdated.

  • 04

    Reads the link you paste

    A link in the question is read before research starts. That covers web pages, PDFs cited by page, YouTube videos cited by timestamp, arXiv and PubMed papers in full text, and GitHub repositories. Filters set a time range and the sites to include or leave out.

  • 05

    Five depths, and a plan first

    Quick, Research, Deep, Academic and Exhaustive are each priced before they run and refunded in full if they fail. The three deepest draft a plan of questions and sections first, which can be edited before the research begins.

  • 06

    Subjects on a schedule

    Standing questions run daily, weekly or monthly, and a compare view shows what changed between two runs. A bell in the header says when a run has finished.

  • 07

    Projects

    One piece of work keeps its chats, reports, subjects and files together, with a default brief and shared instructions for everything started inside it.

  • 08

    Incognito sessions

    A private chat stores no chat, message or report, and its uploads are deleted when the session ends.

Screens

Inside Clarixity

  • The Clarixity home screen, an ask box with brief chips above it, the five research depths with their credit prices inside it and an Incognito toggle below it.
  • A Clarixity answer on the capital of Australia, with superscript citations, a confidence line reading well supported by 16 independent sources, and a rail of 18 sources with quality badges.
  • The Clarixity subjects page, where standing questions are tracked on a schedule.

Architecture

Built end to end

Clarixity is designed, built, deployed and run in house, from the interface down to the server it runs on.

  • Interface

    • Next.js 16
    • React
    • TypeScript
    • Tailwind CSS

  • Research engine

    • Node.js services
    • SearXNG metasearch
    • Playwright reader
    • DeepSeek models

  • Data and payments

    • SQLite with Drizzle ORM
    • Stripe

  • Infrastructure

    • Docker containers
    • Self hosted server
    • Cloudflare
Live

Clarixity is live.

Open to everyone, with an account of your own. Sign up on its site and start in a minute.

Open Clarixity

More products

The rest of the catalogue

Three live AI platforms, architected and shipped end to end. The systems I build for clients and employers are in the engineering record.

The engineering recordAll products