DroFi is a collaboration canvas: upload or draw your project, and it's kept as one blueprint. The difference is simple — that blueprint isn't only for people to read, it's what each teammate's AI works from.
The same three from the home page, one level deeper.
Meeting docs, files, rules, screen flows — the things AI makes up when it doesn't know them. Keep them in the blueprint, and anyone, on any AI, starts every new chat from the same baseline. When the project shifts, nobody has to brief their own AI all over again.
Upload docs, meeting notes, chat logs or GitHub repos and they become the blueprint — then, when you dispatch a task, only the context that fits the role and the task goes to the AI. A developer and a designer get different context, and hunting for files to paste is over.
Guidelines set at the team level become guidelines every AI follows, and no-export documents never reach any AI. Who received what can be checked later, and when something changes, the change reaches everyone's AI automatically.
Same documents, same AI — only the order is new.
Upload the docs, meeting notes, chat logs and GitHub repos you already have, or draw blocks on a blank canvas. Either way you end up with one draft blueprint.
Look at the draft together, adjust the decisions, guidelines and flows, and approve. From then on the blueprint is your team's single source.
Claude, Cursor, ChatGPT, Gemini — each person connects the AI they already use, once. Nobody has to switch tools.
Pick a task, and the context that fits your role and that task goes with it. No hunting for files to paste.
When a decision changes, the change reaches everyone's AI automatically. Explaining it all over again is off your plate.
Lay out decisions, screens, guidelines and flows as blocks and arrows. Several people can edit at the same time and you see who is looking where, live. AI asks about the branches you left out and sketches the next block as a dashed suggestion. A diagram for people, exact context for AI — one screen does both jobs.
Markdown, documents, meeting notes and chat logs — and GitHub repos too — come in as a draft blueprint. Where something is ambiguous, DroFi asks instead of filling it in. Originals stay whole in the document drawer, and the sentences that back a decision are linked to the block they support.
The same blueprint doesn't reach a developer and a designer the same way. When you dispatch a task, only the context that fits the role and the task goes out, so the AI isn't pulled off course by things it doesn't need. What was delivered stays on record.
Commit conventions, tone of voice, things never to do — set them once at the team level and they become guidelines every AI follows. No more results that diverge because everyone prompts differently. Guidelines stop evaporating and start accumulating as a team asset.
Mark the documents that must not leave as no-export. Marked documents never reach any AI — there is no moment where they could slip out by accident. What goes out is up to your team.
Who received what, and through which AI, can be checked later. The path a decision took to get here is kept too, so you can look back at the blueprint as it stood at any earlier point. "Why did we decide it that way?" stops disappearing.
See who used how much, in one place. When your plan's tokens run out, everyone can connect the AI they already pay for and keep working for free. On the Enterprise plan this continues into an audit record of what went out through which AI.
The flow you drew becomes the sequence of work, and AI moves through it one step at a time. Each step carries a "how we know it's done" check, and anything AI proposes changing only reaches the blueprint once a person approves it. You can hand off the work and still hold the standard.
Where it lands depends on your role and scale.
A different AI per person, a different "latest" per tool. One connection puts the whole team on the same context, and if setup feels like a lot, a single copy-paste gets you started.
Every role needs a different slice of context. Send AI only what fits your role, and never miss who changed what.
It's easy to build on stale assumptions when someone else changed a definition. Receive exactly the changes that touch your work and waste less effort.
For people who can't code but want to build with AI — there's a separate track that starts with drawing. DroFi for Vibe Coding →
For those juggling several AIs alone. Draw your standards in one place and the context survives a new conversation, so you never explain from scratch again.
It scales with you. Check the record of what went out to which AI, and self-hosting, enterprise SSO and standard-format export keep you free of lock-in — in control of your own data.
Free for up to 2 editors, no card required. Move up when the team grows.
| Starter Free | Team Popular Pro | Business 10+ people | Max Heavy AI use | Enterprise Custom | |
|---|---|---|---|---|---|
| Price | ₩0 / up to 2 editors | ₩9,900 / editor · month | ₩16,900 / editor · month | ₩24,900 / editor · month | Custom quote |
| Editors | Up to 2 | Up to 10 · viewers free | Up to 50 | Unlimited | By agreement |
| Projects | 2 | 5 | Unlimited | Unlimited | Unlimited |
| AI tokens (per person · month) | 100K (200K for your first 2 weeks) | 1M | 2M | 4M | Unlimited — covered by your company contract |
| And | Canvas for decisions and flows · tailored context delivery · alerts for relevant changes · email, Google and Kakao login | GitHub code import · shared definitions that prevent conflicts · permissions · secure share links · activity log | Unlimited change history · 5 MCP connections | 10 MCP connections · priority support | Context audit log · completion approval gate · SSO/SAML · self-hosting · DPA and security documentation |
On every plan, free or paid, you can connect your own AI and keep working for free once the tokens run out. Full plan details are on the pricing page →
Start free without a card, and move up when the team grows.
Building on your own — DroFi for Vibe Coding → · Evaluating it for an organization — DroFi for Business →