AI assistant for researchers

Research is a long chain of small tasks: finding the right papers, reading them, pulling out what matters, filing notes somewhere you'll find them again, and not missing a submission deadline. MiyoMind collapses that chain into one conversation. You talk to Miyo, the assistant, inside the chat apps you already keep open, and it searches, summarises, files, and reminds without you switching windows.

How does MiyoMind help researchers?

MiyoMind acts as an always-available research companion you message in plain English. It runs live web search and returns answers with citations so you can verify every claim, reads documents and PDFs you send and pulls out the findings, methods, and limitations, writes notes into Notion and other connected tools, and sets one-off or recurring reminders that reach you wherever you chat. It keeps long-term memory of your research focus, so you don't re-explain your project every session.

Here are the workhorse capabilities a researcher leans on most:

• Cited literature search — find recent work on a topic and get answers with links back to sources you can confirm.
• Paper and PDF summarising — send a document and get the abstract, key findings, methodology, and caveats in plain language.
• Note organisation — push structured notes, summaries, and reading lists into Notion via secure OAuth (available once your operator has enabled the connector).
• Reminders that travel — deadline alerts and recurring nudges that fire in WhatsApp, Telegram, Discord, or the dashboard.
• Long-term memory — Miyo remembers your field, current questions, and preferences so context carries between conversations.
• Cross-conversation recall — ask it to pull up what you discussed about a method or paper weeks ago.

What can I ask it to do? (example prompts)

Because you interact in natural language, the easiest way to understand MiyoMind is to see the messages a researcher actually sends. Type these the way you'd text a capable colleague:

Find peer-reviewed work from the last two years on transformer
models for protein folding. Give me the key papers with citations.

I'm attaching a PDF — summarise the methodology and the main
limitations the authors admit to.

Compare how these three papers define "sample efficiency" and
flag where they disagree.

Save a Notion note titled "Lit review — RLHF" with bullet
summaries of everything we covered today.

Remind me every Monday at 9am to update my reading log, and
remind me on June 20 that the conference abstract is due.

What did we say about the dataset bias issue a few weeks ago?

Miyo runs the search, reads the attachment, drafts the note, writes it to Notion, and schedules the reminders — then confirms what it did. You stay in the chat thread the whole time.

A day in the life of a researcher using MiyoMind

Morning, on Telegram: you ask for recent literature on your topic and get a shortlist with citations. You forward two PDFs a collaborator sent and ask for summaries focused on methods. Miyo returns plain-language breakdowns and you ask it to file both into a Notion page for your literature review.

Afternoon, on the web dashboard: you paste a messy paragraph from a draft and ask Miyo to tighten it and check a claim against current sources. It flags one statement that the literature doesn't support and links the papers it found. You set a recurring reminder to back up your data every Friday and a one-off reminder for the grant deadline.

Evening, on WhatsApp: a question pops into your head about a method you read weeks ago. You ask Miyo to recall what you discussed, and it surfaces the earlier conversation and the note it filed — no digging through tabs required.

Is it accurate and safe for academic work?

MiyoMind is built so you can verify what it tells you rather than trust it blindly. Web search results come with citations, so you confirm every factual claim against the source — essential for academic rigour. Treat Miyo as a fast, well-read assistant whose work you check, not an oracle: it can speed up reading and drafting, but you remain responsible for the scholarship.

On privacy and security, every paid user gets a dedicated, sandboxed container with an isolated workspace, no public internet egress, and zero stored external API keys. Your connected integrations and the long-term memories Miyo keeps about your work are encrypted at rest with AES-256-GCM. A 10-layer prompt-injection defence plus output scrubbing runs on every message, which matters when you paste in untrusted text from papers or the web. The free tier runs on a shared direct-agent path instead of a dedicated container.

How is MiyoMind built, and how much does it cost?

MiyoMind runs the open-source OpenClaw agent runtime, a model router called Hermes, and our own orchestration, memory, billing, safety, and routing code on top. It draws on frontier models from OpenAI, Anthropic, Google, xAI, and Alibaba — it is not a thin wrapper around a single model, because the routing, persona, memory, and safety layers are ours. You reach it inside WhatsApp, Telegram, Discord, or the web dashboard at miyomind.com, with nothing extra to install for chat.

Pricing is credit-based, and credits meter your actual model and tool usage rather than a flat per-message fee:

• Free — $0/month, 100 credits every month, no card required, runs on the shared direct-agent path.
• Plus — $14.99/month, 6,000 credits/month, one dedicated sandboxed container.
• Pro — $39.99/month, 18,000 credits/month, one dedicated sandboxed container.
• Top-up packs — 600 for $3, 2,000 for $10, 5,000 for $25, 10,000 for $50, where 1 credit is about $0.005 of value.

For researchers, the practical win is fewer context switches and less lost work: search, reading, note-filing, and deadlines all live in one chat thread you can reach from any device, with citations to keep the work honest and encryption to keep it private.