Every tier gets the whole engine.
Cross-study synthesis, dissonance detection, inflection signals, Ask, evidence chains — identical on every plan. What changes as you move up is how many people share the corpus and how much you put through it, not what the product is allowed to do.
Up to 2 studies per user and 5 users, with the full pipeline and no card on file. If you reach the limit without upgrading, the work you have already done stays intact and readable — you just can't add new studies until you move to a paid plan.
You pay for the people who make research, not the people who read it
Charging per stakeholder is a tax on the exact behaviour you want. Pricing here follows what a person does in the system, and the reading is free.
Creator
The people who run research. They upload, run the pipeline, sit in Triage accepting and rewriting what it proposes, and build the reports. The job shifts from labourer to editor.
- Upload studies and run the analysis pipeline
- Accept, edit and reject proposed insights
- Ask across the whole corpus in scope
- Build and export reports
Explorer
The people who keep asking researchers questions. They get the corpus directly — question in, answer out, citations they can open themselves.
- Ask the corpus in plain language
- Browse published insights and themes
- Open any citation down to the source quote
- Read and share reports
Explorer seats come bundled with Creator seats — no additional charge.
Viewer
Anyone who needs to check the evidence behind a decision — the dev lead, the VP, a stakeholder. No seat purchase, no procurement conversation.
- Read shared reports and claim cards
- Follow evidence chains to their sources
- Browse theme summaries, read-only
Unlimited Viewers on Team and Organization. Three on Professional.
Annual billing saves roughly 20% · Free tier available up to 2 studies per user and 5 users
You go faster.
A researcher with a corporate card can expense this without procurement. Project and study-level permissions let you keep client work isolated — the full intelligence layer, on your terms.
$31/mo billed annually
- Full synthesis — atoms to recommendations
- Cross-study synthesis across your own corpus
- Ask — question everything you have access to
- Inflection signals & dissonance detection
- Claim cards & evidence chains
- Reports: HTML, Markdown, PDF & slides
Your team sees what no individual can.
Connecting your team's data creates cross-study patterns no individual could see. PMs self-serve instead of bottlenecking researchers.
$63/creator/mo billed annually
- Everything in Professional
- Shared team corpus — themes form across people
- Explorer seats bundled — PMs self-serve
- Team workspace with its own scope & taxonomy
- Shared review queue and governance view
- Team admin controls
- In-app comments & threaded annotation
Institutional memory your org has never had.
Org-wide synthesis across hundreds of studies — with every claim traceable to a source timestamp. The kind of research visibility that has never existed at any price.
$119/creator/mo billed annually
- Everything in Team
- Org-wide corpus across every team
- Multiple teams under one organization
- Org-level roles & access administration
- Priority support with a named contact
- SAML SSO & SCIM provisioning
- Immutable audit log
- Executive usage & impact dashboards
- Custom data-retention policies
What Team costs a cross-functional product group
Five researchers, the PMs and designers who ask them things, and everyone else.
Dedicated infrastructure and a longer paper trail
Private vector namespace, custom contracts, invoiced billing, SLAs, data residency, BAA/DPA. Several of these are still being built — the conversation starts with which ones you actually need and when.
The expensive part of research isn't the software
Tool subscriptions are the visible cost. The synthesis hours are the real one, and almost nobody prices them. This is a model of both — move the slider and check it against what you know about your own team.
Matching plan: Team at $79/creator/mo
Doing it by hand
Tool list prices + synthesis labour at a $75/hr blended rate
Umwelt — Team
Full engine, list price, no seat maths
The labour line is an estimate of hours, not a bill you currently receive — most of it is time your team already spends rather than money it already spends. Treat this as an argument about where the hours go, and substitute your own rate if ours is wrong.
And some of it isn't a cost question at all
These aren't in the comparison because no combination of the tools above produces them.
Knowledge freshness alerts
Umwelt tracks when insights are getting stale and flags them — so you're never citing a finding that the market has moved past.
Contradiction detection
Surfaces when findings from different studies conflict with each other. Most teams discover these contradictions only after a bad decision.
Cross-study retrieval
Ask a question that spans multiple studies. The system chains queries automatically and shows exactly where each piece of evidence came from.
Traceable synthesis
Themes and recommendations that keep the atoms they were built from, so any claim can be opened down to the participant and the timestamp.
The things people actually ask
What is the difference between a Creator, an Explorer and a Viewer?
Creators are the researchers: they upload studies, run the analysis pipeline, review and rewrite what it proposes, and build reports. Explorers are the PMs and designers who ask the corpus questions and read the results. Viewers are everyone else — leadership, engineering, stakeholders — who need to check the evidence behind a decision. You pay for Creators. Explorers come bundled with them, and Viewers are free.
Can I keep client or project work separate? I work with several clients.
Yes, and not as a workaround. Studies live inside projects, and access is scoped per project and per study, enforced by row-level security in the database rather than by the interface hiding things. Client A's corpus and Client B's corpus never meet, and an answer can only ever be drawn from the scope you are working in.
Why are Explorer seats bundled instead of priced per person?
Research only pays off when it gets referenced. A per-seat charge on PMs and designers is a tax on exactly the behaviour that makes the corpus worth having, so we took it off. It is also the only way Umwelt gets out of the repository and into the room where decisions are actually made.
Does every tier really get the full analysis pipeline?
Yes. Cross-study synthesis, dissonance detection, inflection signals, Ask, claim cards and evidence chains are on every plan, including the free tier. What changes as you move up is the collaboration surface — a shared corpus on Team and above, org-wide scope on Organization — and how much you can put through it. We do not defend price with feature gates.
What happens if I go over my transcription or storage allocation?
Nothing breaks. We notify you as you approach the limit, and overage is billed at close to our own cost rather than as a penalty. Your usage dashboard shows the projected overage for the current cycle before it closes, so the bill is never the first time you hear about it. Current rates are on your plan page in the app and in any quote we send you.
How does the free tier work, and what happens to my data if I never upgrade?
Up to two studies per user and up to five users, with no card on file and no time limit. Your data does not disappear. If you reach the study limit without upgrading, everything you have already built stays intact and fully readable — you simply cannot add new studies until you move to a paid plan. Umwelt is invite-only during private beta, so the free tier starts with a conversation rather than a signup form.
Why does Organization cost more per Creator than Team?
Because what a Creator can reach grows with the corpus around them. At org scale a single researcher is synthesising across hundreds of studies run by people they have never met, and the patterns available to them are ones that simply do not exist in a five-person corpus. The higher price per Creator reflects a larger reachable corpus, not a longer feature list — the engine is identical.
Can I change plans or reassign seat types later?
Yes. Upgrade, downgrade or reassign seats at any time, prorated. If a solo researcher's company later buys a Team plan, the personal corpus merges into the shared one rather than being re-uploaded, so nothing is stranded on the smaller plan.
Do you train on my research data?
No. Not us, and not the AI providers we call — those run under zero-data-retention terms, so content is processed and returned rather than stored or used for training. Your corpus is never pooled with anyone else's. The security page has the detail.