Your research is yours on every plan.
The full desktop app — runs, logbook, library, papers, compass — is free forever and works entirely on your machine. Paid plans add multi-device continuity, collaboration, and services. Export everything, always, on any plan.
Free
- The complete desktop app — nothing removed
- Unlimited local workspaces, runs & logbook entries
- Automatic run capture with full provenance
- Sealed, tamper-evident logbook
- AI assistant with your own API key (Claude, GPT, Gemini, Mistral)
- Encrypted sync: 2 devices · 1 workspace · 500 MB
- Mobile companion in capture mode
- Full export & transmission packages
- No shared spaces or messaging
Pro
- Everything in Free
- Full mobile companion — runs, comparisons, PDF library
- 10 devices · unlimited synced workspaces · 20 GB
- Project files sync: scripts and data travel with the workspace — a second computer can run the project, and the server doubles as an encrypted off-site backup
- Remote runs: queue an experiment from your laptop, execute it on the machine that has the GPU and the data — results sync back as ordinary runs
- Duo spaces: an encrypted shared space for up to 3 people — a student and supervisor, or co-authors — plus E2EE direct messages with anyone you share a space with
- 150 AI credits/month — hosted assistant, no API key needed (your own key still works, unmetered)
- Qualified timestamps on sealed entries — verifiable proof of when a result was recorded
- Priority support & early features
Lab
- Everything in Pro, for every member
- Teams, member directory & opt-in expertise profiles
- Team & cross-team project spaces with roles
- E2EE messaging: direct messages, group chats, announcements — the lab can never read a member's DMs
- Governance: director's overview & audit trail (metadata only) + progress reports members publish
- Transmission workflows when people leave
- Lab-local AI: the assistant runs on a model hosted on your own hardware (Mistral, for example) — prompts never leave the lab's network; we help you set it up
- Onboarding, support & DPA
- Options: SSO, dedicated instance, self-hosted
Three ways to bring your lab on board
Labs pay by invoice or purchase order — no credit card. Every offer includes deployment help, and the pilot is designed to prove value to the whole team before a longer commitment.
| Offer | Members | Storage | Includes |
|---|---|---|---|
| Lab pilote | 5 | 20 GB | One shared space, onboarding session, email support — 12 months, convertible into a full offer |
| Lab standard | 20 | 100 GB | Teams & directory, roles, audit, DPA, dedicated support hours |
| Lab avancé | 50 | 500 GB | Advanced audit, lab-local AI setup, priority support |
Pricing is quote-based: tell us your team size and constraints and we come back with a written proposal. Options priced in the quote: extra members, SSO, dedicated or self-hosted instance, lab-local AI setup, on-site training. Health data (HDS) is out of scope of standard offers.
The fine print, in plain words
What happens if I stop paying?
Nothing dramatic — ever. Your local app keeps working in full, everything stays exportable, and existing synced data stays downloadable. Quotas simply reapply to new things: extra devices go read-only, new uploads pause above the free storage cap. Nobody is ever locked out of their own research.
Why is the whole desktop app free?
Because trust is the product. ResearchOS is local-first and open formats by construction — Markdown, BibTeX, SQLite, JSON. You pay for continuity across devices, collaboration and services, not for access to your own work.
Is the paid sync more secure than the free sync?
No — security is never a paywall. Every plan gets the same end-to-end encryption: your passphrase derives keys that never leave your devices, and the server only ever stores ciphertext.
How does the AI assistant work on each plan?
On every plan you can plug in your own API key from Anthropic (Claude), OpenAI (GPT), Google (Gemini), or Mistral — you choose the provider and model, and prompts go directly from your machine to that provider, never through our servers. Pro adds 150 hosted credits per month (1 question = 1 credit): pick the "ResearchOS" provider and it works with no key at all — those questions are relayed by our server to the model and immediately discarded, never stored. This is the one exception to the ciphertext-only rule, and it applies only when you explicitly select that provider. During the beta, every account gets 10 hosted credits per month to try it. Labs can go one step further: the assistant runs against a model on the lab's own hardware, so prompts stay on the lab network. And on every plan, any element of your workspace can be individually excluded from the AI.
Can my lab self-host everything?
Yes. The sync server is self-hostable (Docker), and the E2EE design means even your own server never sees plaintext. Managed, dedicated, and on-premise options are on the Lab menu.
Who is behind this?
ResearchOS is built by a PhD student who needed it, for researchers of every kind — and billed under the Tramea name. Beta questions: beta@tramea.org.
Everything is free while we build it with you
Join now, shape the product, and keep 12 months of Pro when plans go live.