AI research pre-processing workflow

Turn one paper into a structured research package your next step can actually use

Mdtero turns supported papers into auditable Markdown, translated reading copies, figures, and reusable bundles for agent-first research work.

⬇ Download Extension ZIP Open demo Log in

Start in the browser, then hand the structured package to your next agent step.

Agent Install

Use Mdtero with the agent you already have

Supported platforms are shown here. When you're ready, open Account, choose a key, and copy your install message there.

Supported platforms

OpenClaw (Recommended)

Open Account, choose a key, and copy a personalized install prompt for OpenClaw.

Claude Code

Open Account, choose a key, and copy a personalized install prompt for Claude Code.

Codex

Open Account, choose a key, and copy a personalized install prompt for Codex.

Gemini CLI

Open Account, choose a key, and copy a personalized install prompt for Gemini CLI.

Create or select a key in Account, then copy the prepared install message and paste it directly into your agent chat.

Open Account

Markdown Clean structure for reading and prompting

Figures Downloaded and linked for reuse

Translation Aligned with the same paper structure

Bundle Ready for your next agent step

Entry Points

Start from the browser extension or your agent

Mdtero has two primary entry modes: capture while reading, or continue inside an agent workflow.

Extension

Browser Extension

Capture supported paper pages into a clean research package without leaving the browser.

Open Edge listing

Agent

Agent Handoff

Send one install instruction, then continue directly into translation, comparison, or review drafting.

Open Account

Workflow

From discovery to literature work in four steps

Capture the paper you actually want

Start from a DOI, a supported paper page, arXiv, or a local retrieval helper and keep the acquisition path explicit.

Parse into a structured research package

Get Markdown, figures, downloadable artifacts, and machine-usable task results instead of a fragile one-off scrape.

Translate, organize, and preserve evidence

Use translation only when needed, while keeping the package auditable and ready for downstream review or synthesis.

Hand it off to your next agent step

Move straight into summarization, review drafting, weekly discovery loops, or your own research automation.

Use Cases

Three research pain points the current product already solves

Mdtero should feel like an entry point into AI research work, not just a parser. These examples map directly onto the current product surfaces.

Weekly Discovery

Pair Semantic Scholar discovery with a reproducible capture loop

Use Semantic Scholar to surface new papers, then feed the shortlisted links into Mdtero so each candidate becomes a consistent package for screening and note taking.

Every Friday, collect this week's promising papers from Semantic Scholar, send the selected links through Mdtero, and return a shortlist with package links and one-line relevance notes.

Translation

Translate a paper without losing its research structure

Instead of pasting a raw PDF into a chat, start from the Markdown package and produce a cleaner translation that still preserves figures, sections, and references.

Take this Mdtero package, translate the paper into Chinese, keep the section structure, and flag any equations, figures, or ambiguous terminology that need manual review.

Literature Review

Draft a literature review from evidence-preserving packages

Once a set of papers is normalized into the same package format, your agent can compare claims, extract methods, and draft a review without fighting format drift.

Compare these Mdtero packages, group them by method and evidence type, then draft a literature review outline with citations back to each paper package.

Pricing

Simple pricing for research reading, translation, and review work

Free

Start with the core workflow.

30 parses / mo
3 translations / mo
Source file downloads

Standard

$3.9 / month

Default for most researchers who want the best value month to month.

120 parses / mo
30 translations / mo
Priority processing

Pro

$7.9 / month

More room for heavier review, comparison, and synthesis work.

300 parses / mo
100 translations / mo
Priority processing

Translation Add-on

$2 / pack

Add this when you only need more translations without changing plans.

20 translations / pack
Translation only
Best for topping up after plan quota

API pay-as-you-go stays available for flexible product and automation use, but plans remain the default path for most researchers.

FAQ

What researchers usually ask before they commit a workflow to it

What is supported right now?

Today Mdtero already supports supported paper-page capture, DOI-led parsing, arXiv-friendly flows, translation, and agent/API handoff on the same product surface.

Do I have to change my whole research workflow to use it?

No. Mdtero is designed as a preprocessing layer that fits into your current reading, note-taking, API, and agent workflows.

Is translation mandatory?

No. Translation is optional. The core product is the structured package that makes the next research step easier.

How does payment work at launch?

Plans cover routine usage first with included monthly quota. Translation add-ons top up extra translation volume, balance remains available for flexible use, and API pay-as-you-go is there for product or automation workflows.

I'm new here — what is an .md file?

An .md file is a Markdown file: plain text with lightweight formatting for headings, lists, tables, links, and images. Mdtero uses it because both humans and agents can read it cleanly. If you want the easiest viewer, open it in Typora; VS Code and many note apps also work well.

My OpenClaw runs on a server and cannot use a campus IP. What should I do?

Use a split workflow. Keep Elsevier or ScienceDirect acquisition on a machine that really has the right local network conditions — usually your own computer with the Mdtero local helper or browser extension. Then send the resulting Markdown or bundle to your server-hosted OpenClaw.