AiHummerAiHummer AiHummer
Start free Sign inMy account

AI ASSISTANT · FOR RESEARCHERS

You have to find what you read all over again.
Your library starts answering questions.

Telegram, WhatsApp, VK, MAX. You hand it your papers, extracts and notes — then ask who had the idea you need. The answer comes back with the file and the page, not with an invented citation.

This works from your library, not from a search of academic databases. It doesn’t go into external publication databases, doesn’t run your literature search and doesn’t write the text for you. What isn’t in your files, it tells you so.

1 590 ₽Free

10 days. After that — 1 300 ₽ per month when billed annually.

Get startedSee the live demo
Moscow
Innovation
Cluster
My libraryonline
Who criticized this sampling method?
Searching your sources…
In two of your sources: the 2021 review, p. 14, and the preprint, p. 3
IN THE LIBRARYsources · 214extracts · 60
Pull together the extracts on this topic
Done — the file is above: 19 extracts with citations
EXTRACTS ON THE TOPICTable, 19 rowsSent as a file into this same chat

Fits into what you already use every day

Full integrations catalog →Connect your own →

Services in this ribbon: 34. Which of them are open depends on your plan. View all plans →

  • Telegram
  • Pocket Agent
  • SIP
  • WhatsApp
  • Google Calendar
  • Yandex Disk
  • Gmail
  • Google Tasks
  • Microsoft To Do
  • OneDrive
  • Signal
  • Todoist
  • Telegram
  • Pocket Agent
  • SIP
  • WhatsApp
  • Google Calendar
  • Yandex Disk
  • Gmail
  • Google Tasks
  • Microsoft To Do
  • OneDrive
  • Signal
  • Todoist
  • Einstein
  • MAX
  • Avito
  • Email
  • Website widget
  • Битрикс24 CRM
  • Google Contacts / People
  • Matrix
  • Moodle
  • Outlook Calendar
  • Slack
  • Einstein
  • MAX
  • Avito
  • Email
  • Website widget
  • Битрикс24 CRM
  • Google Contacts / People
  • Matrix
  • Moodle
  • Outlook Calendar
  • Slack
  • Mobile app
  • Bitrix24
  • VK
  • Telephony
  • Obsidian
  • Discord
  • Google Drive
  • Microsoft Teams
  • Notion
  • Outlook Mail
  • Telfin
  • Mobile app
  • Bitrix24
  • VK
  • Telephony
  • Obsidian
  • Discord
  • Google Drive
  • Microsoft Teams
  • Notion
  • Outlook Mail
  • Telfin

Where a researcher’s working day goes

10:15
which paper that was in
12:40
downloaded the same thing a third time
15:20
the extract is here, the source is gone
18:05
re-reading what I have already read

The library grows faster than it gets read. The expensive part isn’t reading — it’s re-reading: you have seen the idea before, but where exactly won’t come back to you.

The same questions about the methodology

PDFp. 4
Reply to the co-author
LICo-authorL. IlyinaThursday, 21:40
How was the sample drawn?from your protocol
Selection criterionage and length of service
Who was excludedstated on p. 4
taken from “Selection protocol”, p. 4 on Yandex Disk — the file and the page are named
Today

Every new person asks the same thing

The methodology is written down, but not everyone reads it. The question lands in the chat in the evening, and you go hunting again for the file it was in.

With the agent

Answers from your protocol and names the file

It pulls up the right passage in your materials and shows the source — the file and the page. The answer can be checked at once, and hunting for it is over.

ObsidianPrep sheet
Memo on the methodology
Frequent questions of the weekcollected
Sampling, exclusions, scalesnine items
Every item has a sourcefile and page
  • note in Obsidian
  • as a file in the chat
collected from your files on Disk and notes in Obsidian; what is not in them was not added
Today

You explain the methodology from scratch every time

You have written that answer already — in an email, in a chat, in a comment on a draft. Finding that text takes longer than typing a new one.

With the agent

Folds the repeat questions into one memo

It gathers the answers from your own materials and puts the memo into Obsidian as a note. After that it is enough to send it in reply.

Three works on your library

Reply — login only. Then he collects sources from Disk and Obsidian, puts the analysis back into your notes database and gathers you for a calendar call: these are three different jobs, and he does them himself.

PDFp. 14
An answer with sources
Review of 2021 — PDF on Diskp. 14
Preprint, folder “selection”p. 3
Your note in ObsidianMarch
Yandex Disk · Obsidian — the answer names the file and the page, not a paraphrase
Today

Your library has scattered across three storage places

PDFs on Drive, notes in Obsidian, fragments in Notion. To put together a single answer you have to open all of them.

With the agent

Collects sources from three places

PDF on Drive, notes in Obsidian, pages in Notion — one answer, and in it is a file with a page.

XLSX19 rows
Extracts on the topic
Pull together the extracts on samplingyou, in chat
A summary by source19 rows
The same extractas a note in Obsidian
  • Source
  • Page
  • Extract
  • Comment
as a file in the chat and as a separate note in Obsidian or a page in Notion
Today

The analysis settles in the chat and stays there

Extracts pile up in the chat. Moving them into your working note base is a manual job — which means almost never.

With the agent

Puts the analysis into your database

The summary of statements is sent as a file to the chat and as a separate note in Obsidian or a page in Notion.

Thu15:00
Before the call
Call with your co-authorThu 15:00
Pulled up your notes on the topic4 of them
Still unresolvedin one list
Google Calendar · Obsidian
Today

You prepare for the call in the last five minutes

What was discussed last time and what was left unresolved comes back to you only during the meeting.

With the agent

Prepares you for the call

Looks at the calendar, picks up your notes on the topic of the meeting and puts the unresolved items into one list.

One day you don’t lose is worth more than a month of the assistant

Try it on your own library →
8 hours
Lost every week

In the example: hunting for which paper an idea was in; downloading the same thing a third time; re-reading what you have read because an extract lost its source.

1 300 ₽
Assistant for a month

When billed annually. Answers in your own messenger, and from your phone too.

What gets counted here is hours, not revenue

You have already gathered the sources — they just don’t answer questions. No use to you? Turn it off in one click.

What you no longer need to remember and what you don’t need to look for

Reads where you store it

Yandex Disk, Google Disc, Obsidian, Notion — sources and notes together.

Answers right away

In the same messenger where you already write to your co-authors.

Writes back to the database

The summary of statements is stored as a note in Obsidian or a page in Notion.

Never invents a citation

What isn’t in your files, it says so — instead of inventing a source.

Live in an evening, no developer needed

Start today →
5 minutes

You connect a channel

Telegram, WhatsApp or the website widget — by the instructions in the dashboard, no developer.

Telegram WhatsApp Widget
one evening

You upload your documents

What is not in your materials it says so, rather than composing it.

PDF Price List 2026DOC Policy… Warranty
in the morning

Answers when you ask

You see every conversation in the dashboard and in the app. Don’t like an answer? Fix it as plain text.

4 chats overnight

What it’s worth to stop re-reading what you have read

1 590 ₽/mo

or RUB 15,900 per year

  • 1 AI agent
  • Telegram, WhatsApp, VK, MAX
  • Mobile app
  • Your papers, preprints and extracts
  • Extracts and tables as a file, straight into the chat
  • A memory of what you worked through
Get started

The trial period is 10 days.

Need your own server?

How the install works →

Your own server or a virtual machine and PostgreSQL — no Docker needed. It installs with a single command, everything else is set up in the web interface.

  • Installs on Linux — a tarball and systemd, updated with one command
  • Conversations and documents stay with you and never leave
  • Runs on a local model — it can work with no internet at all
See how it installs
the Community plan
0 ₽

Forever, no card. Telegram, web chat and the website widget, one agent, your own server. The rest of the channels come with the paid plans.

Get it free

Questions

Does it search academic databases?

No, and it is worth knowing before you pay. It works from YOUR library — the files you gave it. It never goes into external publication databases and won’t run your literature search. What it does instead is never invent a citation, and in research that is worth more than breadth.

Can it make a source up?

It answers only from your files, and with every answer it names the file and the page — you open it and check. What the library doesn’t have, it says outright: “I don’t have that.” That is a deliberate limit, not an unfinished feature.

What format should I upload sources in?

It’s easier not to download, but to connect: a folder with sources on Yandex Disk or Google Drive, your Obsidian storage, Notion workspace. Then the new PDF, placed in the folder, responds to you without separate work. You can also send the file directly to correspondence. There is no need to rename or organize anything into folders.

Will it write the paper for me?

No. It helps you find and collect what you have already read: answers sources, collects extracts into a table and puts the summary as a note in Obsidian or a page in Notion, prepares you for a call on the calendar. You write the text — and you are also responsible for it.

What is the file it sends?

A table — CSV or XLSX — that it builds itself and sends as an attachment right into the chat. That is how extracts go out with their sources and pages, along with what you have read on a topic and what is still left. The file is ready to write from.

Does it remember what we already went through?

Yes. It brings up your earlier messages and answers with them in mind: which method you rejected and why, what you built on. The limit is an honest one — if you settled it with a co-author by voice and it never reached the chat, it doesn’t know.

Will my unpublished work get out?

No. Your files are used only to answer you. You connect the model yourself — it can be a local one, on your own hardware; then nothing leaves at all. If that isn’t enough, the same product installs on your own server.

Do I need a separate assistant for every topic?

This plan includes one assistant, and the whole library sits with it in one place — usually an advantage: that way it sees the links between topics. If you want the topics kept apart, the next cloud tier has seats for that; how many is written on the pricing page.

What has to be done before the first answer?

Connecting a model is the first step in the dashboard, and you do it once. Until a model is connected the assistant stays silent: the thinking happens in the model, not on our server. After that you connect a messenger and upload your library.

Can I see it before paying?

Yes. The demo sandbox is open without sign-up — login and password below, it resets every hour.

Don't take our word for it — check for yourself

Open the sandboxStart in your account

This page is about one job. Everything about the whole product — how it works, pricing, channels and security — on the home page.

The same assistant — and not only in research

Same mechanics, different reasons to write: a student makes sense of their own lecture notes, a developer stops holding the whole project in their head, a teacher sheds the fifth explanation of the same thing.