Instead of one chatbot — a meeting of four AI models on your Android phone. They answer independently, a moderator compares them into one summary. Free, no account, your own API keys.
A startup from Ukraine, founded by Rostyslav Chervinskyi.
This page is designed to help you find out whether Molfar System is good and if it is the right choice for you.
Molfar System turns an ordinary chatbot conversation into a structured meeting.
You assign up to four AI participants — each on its own model and its own provider (Google Gemini, NVIDIA NIM, OpenRouter), each with a role and its own documents. Then you ask your question. The participants answer independently and can't see each other's replies, so no model adapts to an opinion already voiced.
The fifth slot belongs to the moderator: it compares every answer and pulls them into a single summary — shared points, disagreements, recommendations. The moderator then builds a follow-up queue on its own: specific questions to specific participants, based on what's still unresolved. You can edit those questions, change the recipient, delete the ones you don't need — or run the next round as it is.
The sixth slot is the secretary. It maintains the project's knowledge base, searches your documents and names the source file, and writes the meeting protocol.
For quick questions there's Gaiduk — a single chat with one model, any role, and up to five documents.
LIKE, not vectors — embeddings are on the roadmapListed in
AI participants
Up to 4, each on its own model and provider
Providers
Google Gemini, NVIDIA NIM, OpenRouter
Roles
23 built-in, prompts editable, custom roles supported
Documents
txt, md, docx, pdf — stored on device
Data storage
Local only, no backend, no accounts
Languages
English, Ukrainian, German, French, Spanish
Requirements
Android 7.0+ and your own free API key
Most AI apps give you one model and one answer. Molfar System gives you a room.
There's no backend: projects, chats and documents stay on the device, and requests go straight from your phone to the provider using your own API keys.
Not because it's better at any single answer — the models are the same ones you'd use elsewhere.
It's for the cases where one answer isn't enough: comparing positions, catching what a single model missed, working through your own documents with several expert roles at once.
Practical differences:
The trade-off: you manage your own API keys.
Two groups.
People who work with documents they can't send to someone else's cloud — lawyers, analysts, auditors, consultants, researchers. Local storage isn't a feature here, it's the reason the app exists.
People curious about multi-model setups who don't want to wire up Python, a terminal and agent frameworks just to try one.
It's a semi-research project, so part of the audience is simply testing where several independent models genuinely beat one — and where the extra complexity isn't worth it.
I'm a solo developer from Ukraine. Seven months ago I wrote the first line of this app on my phone — and never once sat down at a computer. The app, the website, the promo images: all of it was built on a phone.
That constraint shaped everything. No desktop meant no Android Studio, so the stack became Python with Kivy, packaged by python-for-android, with builds running in GitHub Actions instead of on a machine I don't have.
The idea itself came from a habit: asking the same question in three different chat windows and comparing the answers by hand. That's tedious, and worse, once you paste one model's answer into another window it starts agreeing with it. So I built the comparison into the app, and made sure the participants can't see each other.
It's a semi-research project — no monetization, no investors, no company. Part of it works well, part still needs proving.
No backend, no server framework, no cloud database — there's nothing running on our side.
None — and by design. There's no monetization, no accounts and no analytics, so there are no customers to name and no way to know who uses the app.
We have collected here some useful links to help you find out if Molfar System is good.
Check the traffic stats of Molfar System on SimilarWeb. The key metrics to look for are: monthly visits, average visit duration, pages per visit, and traffic by country. Moreoever, check the traffic sources. For example "Direct" traffic is a good sign.
Check the "Domain Rating" of Molfar System on Ahrefs. The domain rating is a measure of the strength of a website's backlink profile on a scale from 0 to 100. It shows the strength of Molfar System's backlink profile compared to the other websites. In most cases a domain rating of 60+ is considered good and 70+ is considered very good.
Check the "Domain Authority" of Molfar System on MOZ. A website's domain authority (DA) is a search engine ranking score that predicts how well a website will rank on search engine result pages (SERPs). It is based on a 100-point logarithmic scale, with higher scores corresponding to a greater likelihood of ranking. This is another useful metric to check if a website is good.
The latest comments about Molfar System on Reddit. This can help you find out how popualr the product is and what people think about it.
Do you know an article comparing Molfar System to other products?
Suggest a link to a post with product alternatives.
Is Molfar System good? This is an informative page that will help you find out. Moreover, you can review and discuss Molfar System here. The primary details have been verified within the last quarter. So they could be considered up to date. If you think we are missing something, please use the means on this page to comment or suggest changes. All reviews and comments are highly encouranged and appreciated as they help everyone in the community to make an informed choice. Please always be kind and objective when evaluating a product and sharing your opinion.
