Agentic deep research platform
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.
50 credits to start. No card needed.
What is retrieval augmented generation?
ResearchingA real query, answered by the live platform.
0
Research depths
From a one credit answer to an exhaustive report
0
Built in briefs
Clone any of them, or write your own
0
Credits to start
No card, no trial countdown
0
Credit per quick answer
You see the cost before you spend it
How it works
Ask once, or define it once and reuse it
A brief is a saved definition of a piece of research. You name the questions that must be answered and the sections the report should contain, and the built in modes are editable starting points rather than a ceiling.
Ask, or open a brief
Type a question for a direct answer, or start from a brief that already knows what a thorough answer looks like in your field.
It reads, then writes
Searches run across several angles of the subject at once. What comes back is ranked, read and only then written up.
Check the working
Each claim is numbered to its source. Follow any one of them, or export the whole report and send it to a colleague.
Inside a run
It searches, reads, then writes
A question is not one search. Clarixity breaks it into the angles that actually have to be covered, reads what comes back from each, and only writes once it has something to cite.
Planning searches
4 angles identified
Reading the web
38 results ranked
Reading your documents
6 passages matched
Writing with citations
17 claims numbered
An illustration of the process, not a live run.
Retrieval augmented generation gives a model a search step before it writes.
Rather than answering from training data alone, it retrieves passages from an external corpus
and conditions the response on them, so each claim can be traced to a document.[1][2]
Sources
Capabilities
Built for work that has to hold up
A research tool is only worth using if you can check it. Everything here is built so the answer can be followed back to what produced it.
Depth you choose
A quick answer when that is all the question needs, and a multi angle report when it does not. The depth is a decision, not a guess.
Every claim traceable
Answers carry numbered citations back to what they were read from. When a search turns up nothing relevant, it says so rather than inventing an answer.
Briefs you define
Name the questions that must be answered and the sections a report must contain. Save it once and run it against any subject.
Research that repeats
Track a subject daily, weekly or monthly under a brief. Because the questions are fixed, you can compare any two runs side by side and see what changed.
Your own sources
Pin a brief to the domains that matter, such as a regulator or a journal, and upload your own PDFs and documents to search alongside the open web.
Any field
Scientific, regulatory, technical, financial or competitive. The engine does not change, only what you have told it thorough means.
Scheduled research
Research that keeps itself up to date
Put a question on a schedule and Clarixity runs it again every day, week or month. Each run lands as a digest, and you can see exactly how the answer moved.
Daily, weekly or monthly
A subject asks the same question on the cadence you choose, at 07:00 in your time zone. The cost of each run is shown before you save it.
Compare any two runs
The compare view lines two runs up section by section and marks each paragraph as added, removed or unchanged.
A digest from every run
Each scheduled run files its report in your digests and sends a notification, so a change reaches you without a search.
EU AI Act enforcement
An illustration of the compare view, not a live run.
Mon 14 Sep
PreviousEnforcement
The AI Office is consulting on the code of practice for general purpose models.
No national authority has yet opened a formal case under the Act.
11 sources cited
Mon 21 Sep
LatestEnforcement
The AI Office is consulting on the code of practice for general purpose models.
Removed. No national authority has yet opened a formal case under the Act.
Added. A first national authority has opened a case under the prohibited practices rules.
14 sources cited
New digest ready
1 added · 1 removed · 1 unchanged
Pricing
Pay for the research you run
Every account starts with 50 free credits. No card needed. A quick answer costs 1 credit and a research run 5, and a run that fails is refunded in full.
Plan
Best valueStarter
A fresh allowance every month. Cancel at any time.
$5a month
100 credits every month
- About 100 quick answers or 20 research runs
- A fresh allowance every month, unused credits do not carry over
- 5 cents a credit
Plan
Plus
More room for regular research across a month.
$15a month
300 credits every month
- About 300 quick answers or 60 research runs
- A fresh allowance every month, unused credits do not carry over
- 5 cents a credit
Plan
Regular
Scheduling, exports and your own sources unlock here.
$20a month
420 credits every month
- About 420 quick answers or 84 research runs
- A fresh allowance every month, unused credits do not carry over
- 4.8 cents a credit
Plan
Team
For a small team researching every day.
$50a month
1,105 credits every month
- About 1105 quick answers or 221 research runs
- A fresh allowance every month, unused credits do not carry over
- 4.5 cents a credit
Plan
Business
The developer API unlocks here, alongside the highest allowance.
$200a month
4,651 credits every month
- About 4651 quick answers or 930 research runs
- A fresh allowance every month, unused credits do not carry over
- 4.3 cents a credit
Prices in US dollars. Payments are handled by PayPal.
In practice
Built for the questions that need checking
The engine does not change between fields. What changes is what you have told it a thorough answer looks like.
Regulatory and policy
Track what a regulator has actually published rather than what was reported about it. Pin a brief to the register, run it as a monthly subject, and compare this month with last.
Scientific literature
Ask what the current evidence says and get an answer where each claim carries the paper it came from. The literature review brief searches scholarly sources first and asks where the papers disagree.
Competitive and market
Follow a set of companies against the same fixed questions, so a change in the answer is a change in the world rather than a change in how you asked.
Technical and engineering
Search documentation, specifications and your own uploaded material together, with every claim numbered to the source it was read from.
Questions
Before you start
A chatbot answers from what it absorbed during training, and you have no way to check any particular sentence. Clarixity searches first, reads what comes back, and numbers each claim to the source it came from. When a search turns up nothing relevant, it says so rather than inventing an answer.
Start with a question
Fifty credits to start, no card needed. Enough to run real research and decide for yourself whether the answers hold up.