Pricing a second brain, honestly
Free tiers are usually a trap disguised as a gift. Here's what we tried to get right about tentus's pricing, and where the real trade-offs actually sit.
Most free tiers in this category are a trap dressed up as generosity: capture as much as you like, then hit a wall the moment retrieval — the entire point of a second brain — actually matters. We didn't want to build that trap, so pricing tentus took longer than it should have.
What Free actually gets
Full capture. Full semantic search over what you've captured. Export, always, on every tier — your notes are yours whether or not you're paying, and a second brain that holds your thinking hostage behind a paywall isn't one we'd trust either. What Free doesn't get is the proactive layer (Lumi's nudges, the daily digest) and a cap on how many external integrations you can connect at once. Those are meaningfully more expensive to run, and meaningfully more valuable once you're relying on tentus daily — which is the actual signal for whether upgrading makes sense for you.
Why Business exists as a separate tier from Pro
Business isn't "Pro but more expensive for the sake of it." It's a flat $20/month, single-user tier built around what people who've made tentus a real daily habit actually ask for: essentially unlimited AI actions instead of a monthly cap, and access to Vexa, the visible meeting-bot option for platforms that don't support an invisible native integration. It's priced flat, not per-seat, because tentus is built for one person's thinking, not a team's — Business is for someone who wants more, not a company buying licenses.
The quota question, answered honestly
Research and Composio-routed actions share one combined monthly quota rather than each getting a separate number. That wasn't the plan from day one — it's the result of noticing that stacking a fresh quota onto every new AI-cost feature was heading toward five different numbers a user would have to remember and none of them would mean anything. One pool, one number, is more honest about what you're actually paying for: a bounded amount of the tool doing real work on your behalf, not five artificially separate meters.
What we're still deciding
A full credit-wallet model — buy tokens, spend them across anything, top up when you run low — is the more flexible long-term answer and it's not built yet. It's real, deliberate future work, not a placeholder we're avoiding naming. For now, the simpler tiered-quota model is the one that's actually shipped, tested, and honest about its own limits.